Full text of "Mother Tongue Newsletters"

See other formats

TONGU

SLET

THE ASSOCIATION

FOR THE STUDY OF

LANGUAGE I

PREHISTORY

Issue 22, May 1994

CONTENTS

I Comments on Colarusso's Paper "Phyletic Links between Proto-Indo-
European and Proto-Northwest Caucasian": Allan R. Bomhard

II Is Proto-Indo-European Related to Proto-Northwest Caucasian?
Merritt Ruhlen

13 Comment on Colarusso 1994: John D. Bengtson

16 Indo-European and Uralic Tree Names: Lyle Campbell

3 1 Report from the Field: The Tocharians: Karlene Jones-Bley

3 1 On the Genetic Classification of Basque: John D. Bengtson

36 - Letter from Jerry King to Hal Fleming

36 - A Note on Ofo skdlo "Head": Jerry King

37 - A Note on Catawba Weyaline "Chiefs Town": Jerry King

38 Altaic, Germano-European, and Nostratic: The Evidence of Phonetics
and Phonological Systems: TobyD. Griff en

50 IE Laryngeals — Ever Listened to Them?: W. Wilfried Schuhmacher

51 In the Public Media:

50 - Siberian Site Cedes Stone-Age Surprise

51 - Asian Hominids Make a Much Earlier Appearance

52 - Fossils Put New Face on Lucy's Species

53 - Modern Humans Linked to Single Origin

53 - Mammoth Trumpet Reports the Debate about Amerind Dates

54 Obituaries:

54 - Marija Gimbutas

54 - Sherwin Feinhandler

55 - Susan Park

56 Quick Notes and Hints of Things to Come: Harold C. Fleming

58 Letters to the Editor:

58 - Letter from Henry Harpending

60 - Letter from Ypsilanti

60 - Response to Letter from Ypsilanti: Harold C. Fleming

60 All Method, No Content: Reply by Hal Fleming to Letter from Poser

64 ASLIP Business

65 Editorial
65 Miscellanea

MOTHER TONGUE

Issue 22, May 1994

OFFICERS OF ASLIP

(Address appropriate correspondence to each.)

President:

Harold C. Fleming
16 Butman Avenue
Gloucester, MA 01930
U.S.A.

Telephone: (508) 282-0603

Vice President:

Allan R. Bomhard
73 Phillips Street
Boston, MA 02114
U.S.A.

Telephone: (617)227-4923

Secretary:

Anne W. Beaman
P.O. Box 583
Brookline, MA 02146
U.S.A.

BOARD OF DIRECTORS

Ofer Bar-Yosef
Harvard University

Ron Christensen
Entropy Limited

Frederick Gamst
University of Massachusetts

John Hutchison
Boston University

Mark Kaiser

Illinois State University

Mary Ellen Lepionka
Cambridge, MA

Philip Lieberman
Brown University

Daniel McCall
Boston, MA

Roger Wescott
Southbury, CT

COUNCIL OF FELLOWS

Raimo Anttila
UCLA (USA)

Luca Luigi Cavalli-Sforza
Stanford University (USA)

Igor M. Diakonoff
St. Petersburg (Russia)

Aaron Dolgopolsky
University of Haifa (Israel)

Ben Ohiomamhe Elugbe
University oflbadan (Nigeria)

Joseph H. Greenberg
Stanford University (USA)

Carleton T. Hodge
Indiana University (USA)

Dell Hymes

University of Virginia (USA)

Sydney Lamb

Rice University (USA)

Winfred P. Lehmann
University of Texas (USA)

Karl-Heinrich Menges
University of Vienna (Austria)

Colin Renfrew
Cambridge University (UK)

Vitaly Shevoroshkin
University of Michigan (USA)

Sergei Starostin

Academy of Sciences of Russia

(Russia)

MOTHER TONGUE

Issue 22, May 1994

COMMENTS ON COLARUSSO'S
PAPER "PHYLETIC LINKS BETWEEN

PROTO-INDO-EUROPEAN AND
PROTO-NORTHWEST CAUCASIAN"*

ALLAN R. BOMHARD
Boston, MA

INTRODUCTION

In the previous issue of Mother Tongue, John
Colarusso presented evidence for a genetic relationship
between Proto-Indo-European and Proto-Northwest Caucasian.
He discussed the phonology of Proto-Indo-European and
proposed a revised ("fortified") phonemic inventory for Proto-
Indo-European, he listed several grammatical formants
common to both language families, and he presented a number
of lexical parallels, including preverbs, numerals, particles, and
"conventional cognates."

He concluded that there was evidence, albeit
preliminary, for a genetic relationship between Proto-Indo-
European and Proto-Northwest Caucasian, and he posited a
common proto-language, which he named "Proto-Pontic."

In my recently published co-authored book entitled
The Nostratic Macrofamily: A Study in Distant Linguistic
Relationship (Mouton de Gruyter [April 1994], 932 pp.), I
present a considerable amount of evidence for a genetic
relationship between Proto-Indo-European and certain other
language families, to wit, Proto-Afroasiatic (Proto-Afrasian),
Proto-Kartvelian ("South Caucasian"), Proto-Uralic-Yukaghir,
Proto-Elamo-Dravidian, and Proto-Altaic (Mongolian, Turkic,
Tungus, and probably Korean). Following Pedersen (and Illic-
Svityc and Dolgopolsky), I posit a common ancestor named
"Proto-Nostratic." I also list possible cognates found in
Sumerian and note that Nivkh (Gilyak), Chukchi-Kamchatkan,
and Eskimo-Aleut are probably to be included as members of
the Nostratic macrofamily as well.

There is a growing feeling among many of the
scholars studying Nostratic that Proto-Afroasiatic may be a
sister language of Proto-Nostratic rather than a descendant
language, while Proto-Indo-European is seen by Greenberg as
being more closely related to Proto-Uralic-Yukaghir, Proto-
Altaic, Nivkh, Chukchi-Kamchatkan, and Eskimo-Aleut, these
forming a separate subgroup called "Eurasiatic."

However, Indo-European presents some special
problems. On the one hand, its grammatical structure, at least
in its earlier periods, more closely resembles those of its sister
Eurasiatic languages; on the other hand, its phonological
system more closely resembles the phonological systems found
in Proto-Kartvelian and Proto-Afroasiatic, at least when using
the revised Proto-Indo-European phonological system
proposed by Gamkrelidze-Ivanov and Hopper. Moreover,
there are typological problems with every phonological system
proposed to date for Proto-Indo-European — one wonders, for
example, why there are no affricates. This leads me to suspect
that Proto-Indo-European may in fact be a blend of elements
from two (or more?) different languages, 1 as has already been

suggested by several other scholars. But a blend of what? In
footnote 1 in his paper, Colarusso notes that "[t]he amateur
archeologist Geoffrey Bibby suggested in 1961 that PIE was a
Caucasian language that went north and blended with a Finno-
Ugrian tongue." This suggestion merits closer consideration.

In this paper, I would like to discuss how Colarusso' s
theories shed possible light on these and other issues, noting
both the strong points and limitations of his approach, and I
will propose an alternative theory that I believe better fits the
linguistic evidence.

PHONOLOGY

In Chapter 2 of my book, The Nostratic Macrofamily,
I discuss the phonological systems of the various Nostratic
daughter languages, establish sound correspondences, and posit
a phonological system for the Nostratic proto-language, as
follows:

p[ h ]

t[h]

c[-]

tv[h]

t'y

t|[ h ]

tf'

ky[»]

g y

k'y

k[-]

k w[h]

k'w

q[ h ]

q'w
?

r y ly y

i~e

ry~ey
iw~ew

u~o

uy ~oy
uw~ow

ay ~ay
aw~aw

Colarusso (1994:18) sets up the following phonological system
for Proto-Pontic:

t h

c h

X h

q h

MOTHER TONGUE

Issue 22, May 1994

u
o

Though there are many points of agreement between
Colarusso's Proto-Pontic and my Proto-Nostratic, the main
differences are (A) that I do not posit a separate series of
unaspirated voiceless obstruents (his column 2 above), (B) that
he posits neither palatalized nor rounded tectals 2 nor palatalized
alveolars, and (C) that I posit fewer laryngeals.

For Proto-Nostratic, I set up a series of non-
phonemically aspirated voiceless obstruents (my column 1).
There is actually some evidence, however, that two series may
be warranted: (A) aspirated obstruents and (B) unaspirated
obstruents — exactly what Colarusso has set up for Proto-
Pontic. The evidence for this comes from Afroasiatic on the
one hand and from Altaic on the other. For Proto-Afroasiatic,
a separate phoneme */must be posited in addition to a voiceless
bilabial stop */?, and both of these correspond to voiceless
bilabial stops in the other Nostratic daughter languages.
Setting up two series at the Proto-Nostratic level makes it easy
to account for Proto-Afroasiatic *f 9 which would then be the
reflex of an original phonemic voiceless aspirated bilabial stop
*p h . For Proto- Altaic reconstructions, I followed Nicholas

Poppe and John Street, who posit a simple two-way contrast:
(A) non-phonemically aspirated voiceless and (B) voiced
obstruents. However, Illic-Svityc assumed that there was a
three-way contrast in Proto-Altaic (I have not yet seen the
recent book on Altaic by Starostin, Altajskaja problema i
proisxozdenie japonskogo jazyka [The Altaic Problem and the
Origin of the Japanese Language], Moscow: Nauka [1991]):
(A) voiceless aspirated, (B) plain voiceless, and (C) voiced
obstruents. (Initially, the earlier glottalics are assumed to have
become deglottalized and to have merged with the plain
voiceless obstruents in Altaic, while, medially, they are
represented by voiced obstruents.) If Illic-Svityc 's views are
correct, Altaic may provide evidence for positing separate
voiceless aspirated and voiceless unaspirated obstruents in
Proto-Nostratic. Now, if it turns out that Proto-Northwest
Caucasian is in fact related to Proto-Indo-European as
Colarusso has tried to show, and, by implication, to Nostratic,
additional (and clinching) evidence would be provided for
setting up phonemic unaspirated beside phonemic aspirated
obstruents at the Proto-Nostratic level.

Now, let us look at Proto-Indo-European itself.
Colarusso sets up a three-way contrast for his "Fortified PIE":
(A) voiceless aspirated, (B) plain voiced, and (C) glottalized.
Gamkrelidze-Ivanov also set up a three-way contrast: (A)
voiceless (aspirated), (B) voiced (aspirated), and (C)
glottalized. In their system, the feature of aspiration is viewed
as phonemically irrelevant, and the phonemes of (A) and (B)
can be realized either with or without aspiration depending
upon the paradigmatic alternation of root morphemes. They set
up this alternation mainly to account for instances of
Grassmann's Law. However, as pointed out by Brian Joseph in
a paper read before the 1994 meeting of the Linguistic Society

of America, this reconstruction runs into problems in Italic.
Indeed, it will probably turn out that Grassmann's Law should
not be viewed as pan-Indo-European but, rather, as operating
strictly in certain dialect groups. Now, most scholars,
regardless of whether they follow the traditional reconstruction
of Proto-Indo-European or the radical revisions proposed by
Gamkrelidze-Ivanov and Hopper, set up a three-way contrast
for the obstruents — in other words, they do not set up
phonemic unaspirated voiceless beside phonemic aspirated
voiceless obstruents. The main exception is Oswald
Szemerenyi, who has argued consistently that two separate
series should be set up. The fact is that in most instances the
traditional voiceless aspirates can be explained as secondarily
derived. Moreover, the evidence for their existence is restricted
to one or two branches of Indo-European, and the examples
found there are usually explained as developments specific to
these branches. Nonetheless, there have always been a handful
of examples that cannot be explained as secondarily derived.
In light of Colarusso's proposals, the whole question may merit
re-examination. It may turn out that Szemerenyi has been right
all along. Moreover, setting up phonemic aspirated voiceless
beside unaspirated voiceless obstruents for Proto-Indo-
European may eliminate some of the objections that have been
raised against the revisions proposed by Gamkrelidze-Ivanov.
Thus, one can envision a late (Post-Anatolian) Proto-Indo-
European phonological system of the following shape:

P h

b h

(P')

t h

kvh

gyh

k'y

g h

k wh

k w

gwh

k 'w

Tfi

At an earlier period (Pre-Anatolian), the voiced aspirates
(column 3 above) may have been plain voiced, as I have argued
elsewhere. It may be noted in passing that I view the Anatolian
branch as the first to split off from the main speech community.

The palatalized and rounded tectals and palatalized
alveolars that I reconstruct for Proto-Nostratic present special
problems that have not yet been worked out to my satisfaction.
Consequently, I will not comment further on these series in this
paper. The same can be said for the affricates.

It seems to me that Colarusso posits a greater number
of "laryngeal" phonemes for Proto-Indo-European than
required either by internal Indo-European evidence or by
evidence from the other Nostratic daughter languages.
Extremely good and plentiful cognates containing "laryngeals"
can be established between Proto-Indo-European and Proto-
Afroasiatic, and the "laryngeals" are better preserved in the
Afroasiatic branch than in any of the other Nostratic daughter
languages. For Proto-Afroasiatic, only four "laryngeals" are
posited by most scholars, though there is not unanimity here:
(1) ? (glottal stop), (2) h (voiceless laryngeal fricative), (3) h

MOTHER TONGUE

Issue 22, May 1994

(voiceless pharyngeal fricative), and (4) T (voiced pharyngeal
fricative). There may have been rounded "laryngeals" in
Proto-Afroasiatic as well. I would set up the same four
"laryngeals" for Pre-Proto-Indo-European. I assume, however,
that the voiceless and voiced pharyngeals became multiply-
articulated pharyngeal/laryngeals in later Indo-European, and
this is what I have shown in the reconstruction above. This
assumption is made to account for their vowel-coloring
properties. (I would like to note that this is an idea I originally
got from Colarusso after reading his Ph.D. Dissertation
[Harvard University, 1975].) The whole question concerning
the "laryngeals" remains open, though. The quality and
quantity of the cognates that can be established between Proto-
Indo-European and Proto-Northwest Caucasian on the one
hand and between Proto-Nostratic and Proto-Northwest
Caucasian on the other, especially when the Afroasiatic
evidence is brought into consideration, may require that
additional "laryngeal" phonemes be set up for Proto-Nostratic,
or it may show simply that additional "laryngeals" have arisen
in Proto-Northwest Caucasian through developments specific
to that language family.

NOMINAL SUFFIXES

Colarusso presents a series of nominal suffixes that he
claims are common to Proto-Indo-European and Proto-
Northwest Caucasian. For the most part, I find his examples to
be well founded. I have reservations, however, about the
following (the numbers refer to those given in his paper): in
(16), I would not compare the PNWC suffix */-ga/ — it just
does not seem to belong, both for phonological as well as
semantic reasons; (17) should be removed completely — again,
for phonological reasons; in (18), I see no reason for positing
Proto-Pontic */-m/ when */-n/ is implied by both PIE and
PNWC — I would take the */-m/ found in PNWC to be a
separate ending, perhaps comparable to the */-m/ found in the
PIE accusative case ending.

OTHER ENDINGS

Here, Colarusso considers "some other endings, such
as participles, abstracts, cases, and such." All of the examples
presented by Colarusso are convincing. The one comment I
have is the same as I had above, namely, that I would
distinguish between */-m/ and */-n/.

PREVERBS (OLD NOUNS) / PARTICLES

Colarusso lists three preverbs (old nouns) common to
PIE and PNWC and also compares PIE "final *s" with PNWC
old oblique in */-s/. All of these are convincing examples.
Two ([34] and [35]) of the three preverbs have cognates in the
other Nostratic languages.

Colarusso also lists two particles, both of which are
convincing.

VERBAL DESINENCES AND SUFFIXES

Some of the parallels presented by Colarusso are
intriguing and deserve further investigation. Specifically, I
would like to see more about what PNWC might be able to tell
us about the PIE athematic ~ thematic conjugational types.

I am skeptical about the PIE perfects discussed in
(48). The PIE primary, active, present, athematic *-/ discussed
in (50) is usually derived from a deictic meaning "here and
now"; the explanation offered by Colarusso presents a viable
alternative, though this does not necessarily rule out ultimate
derivation from a deictic.

The explanation given by Colarusso in (52) to PIE "s-
movable" is not convincing and should be abandoned, and the
same goes for the personal ending discussed in (53).

In (56), why the meaning "(in) hand"? I see nothing
wrong with simply positing an "independent adverb before the
verb denoting accomplishment of action" — this seems to be
what is implied in both PIE and PNWC.

STEM FORMATION (a la Benveniste)

Colarusso actually presents an alternative explanation
for certain stem patterns to that offered by Benveniste' s theory
of the Indo-European root. As he himself notes: "[f]urther
work in this area promises to reveal some of the more obscure
cognates between these two families as well as to throw light
upon some of the more difficult laryngeal development within
Indo-European history."

CONVENTIONAL COGNATES

Colarusso lists 20 possible cognates. I find the
following to be the most convincing: (66) "sour, caustic
liquid"; (70) "giant"; (75) "two"; (76) "six"; (79) "son, child,
foster child"; (80) "son, nephew"; (81) "to sit (down)". PIE
"fire" (64) and "period of time" (65) find more convincing
explanations when compared with cognates in other Nostratic
languages. But, note that Ubykh has /-fa-/ "to ignite" — is this
really to be derived from PNWC */-p h a-/ "down; to descend",
or is it a separate stem in its own right? In other words, if there
were justification for setting up a distinct PNWC */-p h a-/ "to
ignite", then comparison with the PIE might be possible and
permit us to posit a Proto-Pontic */p h a-x w 9-r/ "fire", which we
could then easily compare with forms in other Nostratic
languages. (76) "six" and (80) "son, nephew" also have
cognates in other Nostratic languages.

ADDITIONAL COGNATES

I would like to propose some additional cognates. I
have taken the following Northwest Caucasian examples
exclusively from Kuipers's A Dictionary of Proto-Circassian
Roots — it is the only work available to me. Now, I realize full
well that Circassian is but one branch of Northwest Caucasian.
Therefore, adjustments may have to be made to the cognates I
am proposing on the basis of evidence from the remaining
branches of Northwest Caucasian. Moreover, I have brought in

MOTHER TONGUE

Issue 22, May 1994

evidence from other Nostratic languages. Proto-Indo-
European reconstructions are in accordance with the revisions
proposed by Gamkrelidze-Ivanov and Hopper.

The following abbreviations will be used: PCirc. =
Proto-Circassian; PN = Proto-Nostratic; PIE = Proto-Indo-
European; PK = Proto-Kartvelian; PAA = Proto-Afroasiatic;
PEC = Proto-East Cushitic; PHEC = Proto-Highland East
Cushitic; PSC = Proto-Southern Cushitic; PU = Proto-Uralic;
PFU = Proto-Finno-Ugrian; PD = Proto-Dravidian; PAlt. =
Proto-Altaic; CT = Common Turkic; S = Sumerian. The "P"
will be omitted when no proto-form is posited but when there
are cognates from that branch of Nostratic. Other language
names will be spelled out in full.

1. PCirc. *p h arx°a "passageway, porch" ~ PN *p[ h ]ar-
/*p[ h ]ar- "to go or pass; to go or pass over or across; to go
forth or out" > PIE *p[ h ]er-/*p[ h ]or-/*p[ h ]r- "to go or pass;
to go or pass over or across", (n.) *p[ h ]r-t[ h ]- "passage,

crossing, way, ford" (cf. Latin porta "gate, entrance",
porticus "a portico, collonnade, arcade, gallery"); AA:
Egyptian prl "to go, to come out, to go forth, to go up, to
ascend", prw "motion, procession, outcome, result", prt
"(ritual) procession"; S par "to go or pass by, to go past".

2. PCirc. *bays "rich" ~ PN *bay-/*bay- "to apportion, to
divide into shares, to distribute, to allot" > PIE *b[ h ]ey-
/*b[ h ]oy- "to give" (found only in Anatolian); PAA *bay-
/*bsy- "to apportion, to separate into equal shares, to
distribute into shares"; PAlt. *baya(n) "rich"; S ba "to give
as a gift or a ration".

3. PCirc. *ms "this", *maw "thither, that", *.../ma "there you
are!" ~ PN *ma-/*m9- (also *mu-/*mo-) demonstrative
stem > PIE *mo- (found vestigially in Celtic); PK *ma-
"this, he"; PFU *mu "other, another"; Alt.: CT (*mu/*mo
>) (nom. sg.) *bu/*bo, (oblique) *mu-n-; Mongolian mon
deictic word serving as a demonstrative pronoun,
adjective, adverb, and copula.

4. PCirc. *m9 negative prefix ~ PN *ma(?)/*m3(?)
negative/prohibitive particle > PIE *me prohibitive
particle; K: Svan (particle of modal negation) mad "no,
not", mam(a) "not", mama "no"; PAA *ma(?)
negative/prohibitive particle.

5. PCirc. *mana "penis" - PN *many-/*m9ny- "to lust after,
to desire passionately, to copulate, to have sexual
intercourse, to beget" > (?) IE: Modern Irish mian
"desire", Welsh mwyn "enjoyment, value; gentle, kind,
dear"; PAA *man-/*m9n- "to lust after, to desire
passionately, to copulate, to have sexual intercourse, to
beget"; PD *man- "to love, to wed, to copulate with; to be
united with, to marry", *mani "penis". PN *many-/*m9ny-
"progenitor, begetter, man, male" (derivative of the
preceding) > PIE *man(u)- "man, begetter, progenitor";

(?) AA: PHEC *man(n)- "man, person; (pi.) people", PEC
(with fossilized feminine suffix) *man-t-/*min-t-
"woman" (note also Bayso man-to "penis", man-tiiti
"vagina", Burji munn-aa "vagina"); PFU *manyty3 "man,
male"; D: (?) Tamil mantar "human beings, male
persons", Naikri mas "man, husband", Parji mafiji "man".

6. PCirc. *nag(a) "bad, evil; to disfigure" ~ PN *nag-/*nsg-
"to strike, to split, to pierce" > PIE *neg[ h ]-/*nog[ h ]- "to
strike, to split, to pierce"; PAA *nag-/*n9g- "to strike, to
split, to pierce".

7. PCirc. *nap h a "face" ~ PN *nap[ h ]-/*n3p[ h ]- "to breathe,
to blow" > PIE *np[ h ]- > (with metathesis) *p[ h ]n-ew-
/*p[ h ]n-ew-/*p[ h ]n-u-, *p[ h ]n-ek[ h ]- "to breathe, to blow";
PAA *naf-/*n9f- "to breathe, to blow" (cf. Hebrew 'acp

"nose, nostril, face", Arabic 'anf "nose, forepart of
anything").

8. PCirc. *k'ak'a "to chirp" ~ PN *k'ak'- "to cackle, to
chatter" > PIE *k'ak'- "to cackle, to chatter"; PK *k'ak'a-
n- "to cackle"; PAA *k'ak'- "to cackle, to make a noise";
PD *kak- "to laugh".

9. PCirc. *q'°at h a "to tell, to report; to announce, to make
known" ~ (?) PN *k' w aty[ h ]-/*k' w 9ty[ h ]- "to say, to speak,
to call" > PIE *k' w et[ h ]-/*k' w ot[ h ]- "to say, to speak, to
call"; PFU *kuty3- "to call, to summon".

10. PCirc. *q:ara "black" ~ PN *k[ h ]ar-/*k[ h ]3r- "black, dark"
> PIE *k[ h ]er-s-/*k[ h ]r-s- "black, dark"; PD *kar- "black,
dark; to grow black, to darken"; PAlt. *kara "black".

11. PCirc. *ban3 "(children of) family" (semantic develop-
ment as in Mehri and Hebrew cited below) ~ PN *bany-/
*bsny- "to join together, to fit together, to fasten, to twist
together, to form or produce in any way" > PIE
*b[ h ]end[ h ]-/*b[ h ]ond[ h ]-/*b[ h ]nd[ h ]- "to join together, to
fit together, to fasten, to twist together, to form or produce
in any way"; PAA *ban-/*b9n- "to join together, to fit
together, to fasten, to twist together, to form or produce in
any way" (cf. Mehri hs-bon "children"; Hebrew ben "son,
grandson; (pi.) children [both sons and daughters]"); PD
*pan- "to make, to do, to produce, to build".

12. PCirc. *k'°ad(a) "to disappear, to get lost, to perish" - PN
*k' w ad-/*k' w 9d- "to strike, to wound, to hurt, to slay" >
PIE *k' w ed[ h ]-/*k' w od[ h ]- "to strike, to wound, to hurt, to
slay" (cf. Lithuanian gendu, gesti "to go out, to die out, to
become dim"); K: Georgian k'vd- "to die"; PD *kutt- "to
strike, to knock, to pound".

13. PCirc. *t'aw3 "to bump (one's head)" ~ PN *t'aw-/*t'aw-

MOTHER TONGUE

Issue 22, May 1994

"to hit, to strike" > PIE *t'ew-/*t'ow-/*t'u- "to hit, to
strike"; PAA *t'aw-/*t'ow- "to hit, to strike"; S du 7 "to
butt, to gore".

14. PCirc. *wasa "price" ~ PN *wus-/*wos- "to trade, to deal"
> PIE *wes-/*wos- "to trade, to deal", (n.) *wes-no-m
"price" (cf. Sanskrit vasna-m "price, value"); PFU *wosa
"trade, commerce".

15. PCirc. *Aaha "rivulet" ~ PN *lali-/*ton- "to make flow, to
pour, to moisten, to wet" > PIE *lefih- [*larih-] (extended
form *lerih-w/u- [*larlh-w/u-]) "to pour, to pour out
(liquids)", (n.) *lehh-no-s [*larih-no-s] (> *la-no-s),
*lerih-mo-s [*larih-mo-s] (> *la-mo-s) "anything that

contains water or liquid: puddle, trough, tub, vat, etc." (cf.
Greek Xx\v6q "anything shaped like a tub or trough: a
wine-vat, a trough [for watering cattle], a watering-place",
Xf|ur| "rheum"; Latin lama "marsh, puddle"; Old Icelandic
Ion "lagoon, inlet"; Latvian lans "puddle", lama "puddle");
PAA *lafi-/*l3fi- "to make flow, to pour, to moisten, to
wet"; S lah "to wash, to clean; (n.) laundry, wash".

16. PCirc. *gal(a) "to slip / to (slip and) fall" ~ (?) PN *gyil-
/*gvel- "to glide, to slip, to slide" > PIE *g[ h ]ley-/*g[ h ]loy-
/*g[ h ]li- "to glide, to slip, to slide"; PAA *gyal-/*gyal- "to
glide, to slip, to slide"; PFU *kil3 (*kiil3) "smooth,
slippery".

17. PCirc. *warda "high-born", *warq:a "nobleman" - PN
*war-/*wsr- "to raise, to elevate; to grow, to increase; (n.)
uppermost, highest, or topmost part" > PIE *wer-d[ h ]-
/*wor-d[ h ]-/*wr-d[ h ]- "to raise, to elevate; to grow, to
increase", (adj.) *word[ h ]-o-s "grown, full-grown, tall,
upright", *wrd[ h ]- and *wrHd[ h ]- (> *wrd[ h ]-) "raised,
upright, tall", *wers-/*wors-/*wrs- and *werk[ h ]s-
/*work[ h ]s-/*wrk[ h ]s- "uppermost, highest, or topmost
part"; AA: Egyptian wr "great, important; much, many;
eldest", wrr "to be great, to make great, to increase, to
grow large", wr "greatness; great one, chief; PFU *wara
"(wooded) hill or mountain"; PD *varay "mountain,
peak", *var "length, elongation; height, straightness".

18. PCirc. *waAa "to totter, to undulate" ~ PN *waly-/waly-
"to turn, to roll, to revolve" > PIE *wel-/*wol-/*wl- "to
turn, to roll, to revolve"; PAA *wal-/*wal- "to revolve";
PD *val- "to surround, to walk around, to walk in a circle".

19. PCirc. *wala§a "to pound, to crush", *wslaw9 "to get
tired" ~ PN *wal-/*wsl- "to crush, to grind, to wear out; to
be worn out, weak; to fade, to wither, to waste away" >
PIE *wel-/*wol-/*wl- "to crush, to grind, to wear out; to

be worn out, weak; to fade, to wither, to waste away"; PD
*val- "to be tired; to become emaciated, thin; to ache", *ol-

"to grow weak or faint; to become reduced, slender, thin,
emaciated"; PAlt. *61- "to be weak from hunger, to wither,
to fade, to starve to death". PN *wal-/*w9l- "to strike, to
wound, to destroy" (probably identical to the preceding) >
PIE *wel-/*wol-/*wl- "to strike, to wound"; PD *vel- "to

conquer, to overcome, to subdue, to destroy".

20. PCirc. *k'ana "piece, lump" ~ (?) PN *k'yun-/*k'y n- "to

bend or fold together, to crack, to split, to divide" > PIE
(*k'en-/*k'on-/)*k'n- "to bend or fold together, to crack,
to split, to divide", (n.) *k'onk'o-s "lump", *k'onk'ulo-s
"lump, ball, bundle", *k'noHt'o-s (> *k'not'o-s) "lump,
knot", *k'nut[ h ]o-s "lump, knot"; PK *k'on- "to tie or bind

together" (cf. Georgian k'on-a "bundle, bunch"); PAA
*k'yan-/*k'y9n- "to bend or fold together, to crack, to split,

to divide".

21. PCirc. *t:aXa "to splash, to threaten; to shake (fist), to
wave threateningly; to rattle (saber)" ~ PN *t[ h ]aly-
/*t[ h ]aly- "to push, to thrust" > PIE *t[ h ]el- "to push, to
shove, to strike, to hit" (Pokorny 1959:1062 *telegh- "to
hit" [?], *telek- "to push, to shove, to hit"); PAA *t[ h ]al-
/*t[ h ]al- "to push, to thrust"; PFU *toly3- "to push, to
shove, to thrust"; PD *tal- "to push, to shove, to thrust, to
press through".

22. PCirc. *baA3 "to hide" (< "to spread out, to cover over")

- PN *bul-/*bol- "to swell, to spread out, to overflow, to
puff up, to inflate" > PIE *b[ h ]el-/*b[ h ]ol-/*b[ h ]l- "to
swell, to puff up, to inflate, to expand, to bubble up, to
overflow"; K: Georgian *blom- "multitude" in (adv.)
blomad "in a crowd, mass, mob, multitude"; PAA *bal-/
*bsl- "to swell, to expand, to spread out, to overflow" (cf.
PEC *balba[a]l- ["to spread out, to cover" >] "curtain,
canopy, covering, etc."); S bul "to blow, to breathe, to
puff.

23. PCirc. *t'9A9 "to lie down" ~ PN *t'al-/t'3l- "to stretch, to
extend" > PIE (*t'el-/*t'ol-/*t'l- "to stretch, to extend, to
lengthen":) Tl-H-gp]- "long", *t'l-eE-g[ h ]- (> *t'l-e-
g[ h ]-) "to stretch, to extend, to lengthen"; PAA *t'al-/t'3l-
"to stretch, to extend"; S dalla "to widen, to stretch, to
extend, to enlarge". For the semantics, cf. Old Church
Slavic klado, klasti "to lie down; to lay, to put" vs.
Lithuanian kloju, kloti "to lay out, to spread out".

24. PCirc. *q:ara "stony land, rock" - PN *k[ h ]ar- "hard,
strong, firm" > PIE *k[ h ]ar- "hard, strong, firm" (cf.
Sanskrit karkara- "hard, firm", also "stone, limestone");
PAA *k[ h ]ar- "hard, dry"; PD *karu "strength, power".

25. PCirc. *mark'°a "mulberry, blackberry" ~ PN *mur-/
*mor- "mulberry, blackberry" > PIE *mor- "mulberry,
blackberry"; AA: Egyptian mr "mulberry-tree"; PU

MOTHER TONGUE

Issue 22, May 1994

*mura "Rubus chamaemorus" (note also PFU *marya
"berry").

26. PCirc. *hat:a "father" ~ PN *?at[ h ](t[ h ])-/*?3t[ h ](t[ h ])-

"father" > PIE *?at[ h ](t[ h ])- "father"; AA: Egyptian it

"father"; PED *atta "father"; PA *etike(y) "older male
relative".

27. PCirc. *hana "mother" ~ PN *?any- "mother, aunt" > PIE
*?ano-s "mother"; AA: PSC *?arj- "father's sister"; PU
*anya "mother, aunt"; PD *ann- "mother, a woman,
father's sister".

28. PCirc. *Aa "leg, foot" (derivatives include: *Xak h a
"strong, firm, staunch [of person]" [contains *Aa "leg" and
an unclear element]; *Xaq:°a "leg"; *Aas9 "on foot,
unmounted"; *Aamb9 "footprint"; *Aac h 'a "lame"; *Aag°3
"footsole, floor, valley") ~ PN *lak[ h ]-/*bk[h]- "leg, foot"
> PIE *lak[ h ]- "leg, foot"; PAA *lak[ h ]-/*tok[ h ]- "leg,
foot".

29. PCirc. *n(a) "mother" (derivative: *n[a]wa "old woman,
granny") ~ (?) PN *nat'- "woman, female relative" > AA:
PSC *nat'a "woman"; PU *nata "sister-in-law, younger
brother of the husband or the wife"; PD *nat- "husband's
sister, sister-in-law".

30. PCirc. *nart h a "Nart" (mythical hero) ~ PN *nir-/*ner- "to
be strong, manly, virile" > PIE *ner- "to be strong, manly,
virile; (n.) man, hero", *ner-t[ h ]o-s "manly; manliness,
virture, strength", *ner-yo-s "manly; man"; AA: Egyptian
nr "to be strong, mighty", nrw "power, strength, victory,
valor, mighty one"; S ner "prince".

31. PCirc. *wala "cloud" ~ PN *wal-/*wsl- "to flow, to wet,
to moisten" > PIE *welk[ h ]-/*wolk[ h ]-, *welk'-/*wolk'-/
*wlk'-, *welg[ h ]-/*wolg[ h ]-/*wlg[ h ]- "to wet, to moisten",
(adj.) *wolg[ h ]o-s "lukewarm, damp, moist" (cf. Old
English weolcen, wolcen "cloud", German Wolke
"cloud"); AA: Arabic waliha-t "well-watered, rich in
vegetation"; PD *ol- "to flow", *val- "to drizzle, to drip".

32. PCirc. *was hf 9 "ax" (cf. *w[a] "to beat, to strike, to shoot,
to sting, to attack; to smooth [a threshing floor]";
derivatives include: *wab[a] "to trample, to pound, to
crush; to slander"; *wap'a "to cut open [an animal], to
skin"; *wot h a "to trample, to knead clay with the feet";
*w9t:[a] "to pound, to thresh [sunflower seeds, maize], to
beat [egg], to knead, to break off; *wak'9 "to kill";
*wat:a "hammer") ~ PN *wasy-/*wasy- "to beat, to strike,

to crush, to grind, to pound, to wear out; to be or become
worn out, tired, weary, fatigued, exhausted" > PIE *wes-
"to crush, to grind, to pound, to wear out; to wither, to

fade, to rot, to waste away"; AA: Egyptian wss "to crush,
to pound", wsm "to slay, to crush, to chop up, to split, to
pound together"; PFP *wasy& "to be or become tired,

weary, fatigued, exhausted"; PD *vec- "to grow tired,
fatigued, weary", *vac- "to cut off, to chip off, to smooth
by chipping, to shave down", *vacci "adze, scraper".
Buck (1949:561-562, §9.25 ax) notes that words for "ax"
are typically derived from various roots for "cut" or
"strike", and this is the derivation that I assume for both
PCirc. *was h 'a "ax" and PD *vacci "adze, scraper".

33. PCirc. *k:at h a "sword" ~ PN *k[ h ]at[ h ]-/*k[ h ]st[ h ]- "to
beat, to strike, to fight" > PIE *k[ h ]at[ h ]- "to fight"; AA:

Egyptian ktkt "to beat, to strike"; PD *kat- "to be angry
with; (n.) anger, wrath", *kat- "to kill, to murder; to cut, to
divide; to quarrel, to fight; (n.) fight, battle, war", *katti
"knife, sword, razor", *katk- "to strike down, to cut with
an ax".

In the following additional cognates, only Proto-Circassian and
Proto-Indo-European forms are given — I have been unable to
locate cognates in other Nostratic languages.

34. PCirc. *nsba "belly" ~ PIE (*nebp]-/)*nobp]- "navel"
(Pokorny 1959:314-15 reconstructs [*enebh-], *embh-,
*ombh-, *nobh-, [*nebh- ?], *mbh-). Note here also
Temirgoy naba^'s/banza "navel", Ubykh naba^' "navel".

35. PCirc. *ban(a) "to fight" ~ PIE *b[ h ]en- "to slay, to
wound".

36. PCirc. *mab "sheep" ~ PIE *mel- "wool, woollen
garment".

37. PCirc. *hawa "but" ~ PIE *hew- [*haw-] "that, other" (cf.
Gothic auk "but, also"; Latin au-tem "but, on the other
hand").

38. PCirc. *k h ak h a "to chirr, to laugh derisively, to bleat, to
howl, to shout" ~ PIE *k[ h ]ak[ h ]a exclamation indicating
laughter (onomat.).

39. PCirc. *p:aya "enemy" ~ PIE *p[ h ]e(y/i)- "to hurt, to
harm, to attack" (Pokorny 1959:792-93 reconstructs
*pe[i]-; note also Mann 1984-1987:913 *peio "to hurt, to
harm, to attack, to blame" and 935 *p!io "to hate, to
revile") (cf. Gothic fijands "enemy", Old English feond
"enemy").

40. PCirc. *yat h a "to rage (storm); to become insolent" ~ PIE
*yet[ h ]- "to let fly at something" (Mann 1984-87:444
reconstructs PIE *ieto "to dash, to dart, to proceed, to
strive", *ietis "motion, eagerness, dash, passion").

MOTHER TONGUE

Issue 22, May 1994

41. PCirc. *k'9y3 "to shreik, to howl" ~ PIE *k'e(y/i)- "to
sing, to call, to cry out" (Pokorny 1959:355 reconstructs
*ge[i]-: *go[i]-: *gl-).

42. PCirc. *k'9r9 "thick, dense (of wool, beard, etc.), long (of
hair), high (of grass)" (cf. *k'[a] "to come out, to bud, to
grow") ~ PIE *k'ar-k'9ro-s "mass, pile, heap" (Mann
1984-1987:265 reconstructs *gargaros, -ios).

43. PCirc. *k'ans "knucklebone (used in bone game)" - (?)
PIE *k'enu "knee".

44. PCirc. *k'°asa "to go out (as fire, light); to escape, to run
away, to desert, to elope" - PIE *k' w es- "to extinguish"
(cf. Lithuanian gesti "to go out, to die out, to become
dim").

45. PCirc. *q' aya "to bleat" ~ (?) PIE *k' w ey- "to complain,
to bemoan" (Pokorny 1959:567 reconstructs *g y ei-).

46. PCirc. *waro "wave; turbulent" ~ PIE *wer-/*wor-/*wr-
"to twist, to turn".

47. PCirc. *sama "heap" (< "to gather together, to pile up, to
heap up") ~ PIE *sem-/*som- ("to gather together" >)
"together, together with; one" (cf. Sanskrit sa [< *sm-]
"with, together with, along with", sam "with, together
with, along with, together, altogether", sa-tra "together,
together with", samana-h "meeting, assembly; amorous
union, embrace", samubha-h "heap, collection").

48. PCirc. *k h ani9 "to be insufficient, to lack"
"lacking horns, hornless".

< PIE *k[ h ]em-

49. PCirc. *gaya "smooth (of ice)" ~ PIE *g[ h ]ey- "snow, ice,
winter" (cf. Sanskrit hima-h "snow, frost, hoar-frost,
winter", hemanta-h "winter, the cold season"; Greek %mv

"snow; snow-water, ice-cold water", x 8 W "winter-
weather, cold, frost", x£iu<ov "winter; wintry weather, a
winter storm"). The Northwest Caucasian word may be a
loan from Indo-European.

50. PCirc. *q:°st h a "to smash, to break, to chop" ~ (?) PIE
*k w [ h ]et[ h ]- in *k w [ h ]et[ h ]-wer- "four", assuming semantic
development as follows: *k w [ h ]et[ h ]- "to cut into (equal)
parts", *k w [ h ]et[ h ]-wer- "having equal sides" > "square,
four-sided" > "four" (cf. Sanskrit catvara-m "square,
cross-roads", thematic extension of a neuter noun *catvar,
which has not been preserved; also Latin -quetrus in
triquetrus "three-cornered, triangular" and quadra [with
-dr- from -tr-] "a square"). Though this etymology is
highly speculative, it is not impossible. Clearly, the word
for "four" in PIE is a derivative and is probably a fairly late
creation. Note that the Anatolian languages have quite a
different word for "four": Hittite meu-, Cuneiform

Luwian ma-a-u-wa-, Hieroglyphic Luwian mawa-, Lycian
*mu- in mupmm- "four- fold" (Lycian also has kadr-[nna]

"four" [this is not listed in Melchert's Lycian Lexicon]) —
these may reflect the original PIE word for "four".

THE MAKE-UP OF INDO-EUROPEAN

Introductory Remarks

As noted in the Introduction to this paper, the
hypothetical Indo-European parent language presents some
special problems. To repeat, Proto-Indo-European shares a
number of lexical and grammatical cognates with certain other
language families of northern, central, and eastern Eurasia,
especially Uralic-Yukaghir, and this was the reason that
Greenberg included Indo-European as a member of his
proposed Eurasiatic language family. When one looks at the
phonology of Proto-Indo-European, however, especially when
the revisions proposed by Gamkrelidze-Ivanov and Hopper are
brought in, one is reminded of a Caucasian language — but
certainly not a typical Caucasian language.

Typically, the phonological systems of the indigenous
languages of the Caucasus Mountains region are characterized
by extremely large inventories of consonantal phonemes. For
example, the consonantal system of Ubykh contains 80
consonantal phonemes (or 83, when one considers the velar
triad found in words of foreign origin [cf. Comrie 1981:202]).
In contrast, the Proto-Indo-European phonological system is a
model of simplicity. (The phonology of Northwest Caucasian
is discussed in great detail by Colarusso in his 1975 Ph.D.
dissertation The Northwest Caucasian Languages: A
Phonological Survey.)

When one looks for simplicity, nothing could be more
straightforward than the phonological system traditionally
posited for Proto-Uralic, which is usually given as follows:

1 2

P t

5
s

m n

In my co-authored book, The Nostratic Macrofamily,
I use a slightly different notation for the phonemes in column
5: ty (= c), 6v (= 5'), sy (= s), ny (= h), ly (= 1'), y (= j) — this is
an adaptation of the transcription proposed by Gyula Decsy.
An additional sibilant, namely *s, is sometimes also posited. It
may be noted in passing that I take the phonemes traditionally
transcribed as *8 and *5' to have been *t| (or simply *i) and
*t|y (or simply *iy — a rare phoneme typologically, but found,

for example, in Kabardian) respectively.

The reconstruction of the Proto-Uralic vowels, how-
ever, presents many problems, and there is still no consensus.

MOTHER TONGUE

Issue 22, May 1994

Uralic-Yukaghir

The primeval homeland of the Finno-Ugrians around
3,000 BCE is shown in the above map (this is taken from Hajdii
1975:31). What is important here is that, in terms of both
dating and geography, the Finno-Ugrians were contiguous with
Indo-Europeans, who were located just to the south in the area
north of the Black Sea and extending eastward well beyond the
Caspian Sea, that is, along the whole southern periphery of the
Finno-Ugrian speech area. At this late date, however, the Indo-
European proto-language had already begun to split up into
different dialect groups. 3,000 BCE was the period of contact
between Finno-Ugrians and the Indo-European dialect group
that was developing into Indo-Iranian. The undifferentiated
Indo-European parent language is usually placed at about 4,000
BCE or perhaps a little earlier.

The date at which the unified Uralic proto-language is
thought to have been spoken is usually given as approximately
4,000 BCE, which would make it roughly contemporaneous
with the period of Indo-European unity, while bringing in
Yukaghir pushes this date back another millennium or so.
Pushing the date backward in time also narrows the range of the
speech community and moves its location further to the east.
Decsy (1990:9), for example, places the Uralic proto-language
"in the Forest Zone-Steppe-Border (mainly north of it) between
the Volga Bend in Eastern Russia and the Ob River in Western
Siberia."

Now, during the Wurm glaciation, which peaked
about 20,000-18,000 years ago, most of northern Eurasia was
covered by huge sheets of ice, while treeless steppe-tundra
stretched all the way from the European shores of the Atlantic
Ocean in the west to well beyond the Ural Mountains in the

east. It was not until about 15,000 years ago that the ice sheets
began to recede in earnest. Clearly, most of what is considered
to have been the Uralic(- Yukaghir) homeland was inhospitable
to human habitation at one time. Therefore, I assume that, at a
much earlier time, Pre-Uralic speaking people were located
even further to the east and south and that they moved slowly
westward and northward exploiting the opportunities created
by the receding ice sheets.

Indo-European and Uralic-Yukaghir

Both Indo-European and Uralic-Yukaghir are usually
considered to be members of the Nostratic macrofamily (so, for
example, Illic-Svityc 1971- , Dolgopolsky 1984, and
Bomhard-Kerns 1994), while Greenberg (forthcoming) main-
tains that Indo-European is to be included as a member of the
Eurasiatic language family along with Uralic-Yukaghir, Altaic,
Nivkh (Gilyak), Chukchi-Kamchatkan, and Eskimo-Aleut. In
my co-authored book The Nostratic Macro-family, I accept
Greenberg' s classification and set up a distinct Eurasiatic
subgroup within Nostratic.

The following evidence provides the basis for
considering Indo-European to be genetically related to the
other Eurasiatic languages, especially Uralic-Yukaghir: (A)
Cognates in the system of pronoun stems (cf. Bomhard-Kerns
1994:3-7, Dolgopolsky 1984, Greenberg forthcoming, and,
especially, Kerns 1985:11-51) — after carefully comparing and
analyzing the pronoun stems of Indo-European, Uralic, and
Altaic, Kerns (1985:48) states unequivocally: "The results are
overwhelming. We are forced to conclude that the pronominal
agreements between Indo-European and Uralic, between Uralic
and Altaic, and between Indo-European and Altaic, did not
develop independently, but instead were CAUSED by some
UNIQUE historical circumstance. In short, it is extremely
unlikely that the three pronominal systems could have evolved
independently." (B) Morphological cognates (cf. Bomhard
1988 for a comparison of Indo-European athematic verbal
endings with Uralic verbal endings, Collinder 1934 for Indo-
European and Uralic, and Greenberg forthcoming for all of
Eurasiatic). (C) Lexical cognates (cf. Bomhard-Kerns 1994,
Illic-Svityc 1971- , Joki 1973, and Greenberg forthcoming).

Taken together, the evidence for a genetic relationship
between Indo-European and the other Eurasiatic languages is
much stronger by far than the evidence presented by Colarusso
for a genetic relationship between Indo-European and
Northwest Caucasian. Moreover, Indo-European and Uralic-
Yukaghir appear to be particularly close and may represent a
separate subbranch within Eurasiatic. For the sake of
argument, I will assume this to have been the case — this sub-
branch may be called simply "Indo-Uralic".

I have proposed (Bomhard-Kerns 1994:35) that Proto-
Eurasiatic be dated to about 7,000 or 8,000 BCE, though a little
earlier is also possible. I would place Proto-Indo-Uralic to
about the same time period, that is, 7,000 BCE, and I would
locate Proto-Indo-Uralic in Central Asia, as Nichols (1993) has
done for what she calls "Pre-Indo-European". I believe that
Nichols (1993) is correct in seeing a westward spread of Pre-
Indo-European across the steppes, though I would use the term

MOTHER TONGUE

Issue 22, May 1994

"Indo-Uralic" instead for the initial period of westward move-
ment. However, as Indo-Uralic continued its westward move-
ment, I assume that it split into main two dialect groups, with
what was to become Uralic-Yukaghir following a more
northerly route and with what was to become Indo-European
following a more central route straight across the steppes. It is
this southern dialect that may be called "Pre-Indo-European".

One final point may be made here: I believe that the
Proto-Indo-Uralic phonological system had a voicing contrast
in obstruents, and this was preserved in Pre- and Proto-Indo-
European. It was only in Proto-Uralic-Yukaghir that the
voicing contrast was eliminated.

Indo-European and Caucasian

The westward spread of Pre-Indo-European
eventually brought it to the shores of the Black Sea at 5,000
BCE or thereabouts. Though it is not known what language or
languages were spoken in this area before the arrival of the Pre-
Indo-European-speaking people, it is known that the Pre-Indo-
Europeans were not the first inhabitants of the area. I can
envision at least three possibilities: (A) the language or
languages were related to those of the Caucasus Mountains; (B)
Pre-Uralic-Yukaghir-speaking people may have arrived in the
area first and have been there when the Pre-Indo-Europeans
arrived; (C) one or more unknown languages were spoken
there. Scenario (B) is highly unlikely on both linguistic and
archeological grounds. That leaves us with scenarios (A) and
(C) as viable alternatives.

According to Villar (1991:15), the above map shows the homeland of
Indo-European-speaking people at about 5,000 - 4,500 BCE, while the
stroked area above the Caspian Sea indicates the earliest probable
location of the Indo-Europeans.

We can now return to the suggestion made by
Colarusso (1994:8) that "PIE and PNWC are genetically
related at a phyletic level". In view of the strong evidence for
a genetic relationship between Proto-Indo-European and other
Eurasiatic (and Nostratic) languages, Colarusso 's proposal
seems unlikely. However, what Colarusso has done is that he
has rightly identified Caucasian elements in Indo-European.

I will venture a guess that, when the Pre-Indo-
Europeans arrived on the shores of the Black Sea, they met and
mingled with Caucasian-speaking people (scenario [A] above).

It was this contact that eventually produced Proto-Indo-
European, with its greatly simplified Caucasian-looking
phonological system and with its grammatical structure and
system of pronominal stems bearing strong resemblances to
what is found in Uralic. While maintaining contact with what
was to become Uralic on the one hand and with the languages
of the Caucasus on the other, Indo-European gradually
assumed its own identity distinct from both Uralic and
Caucasian.

Scenario (C) above is also a possibility, in which case
Indo-European would have replaced whatever language or
languages were spoken in the area. This means that there is
practically no way to determine what that language or those
languages were, except, perhaps, through the analysis of the
Indo-European lexicon to try to find possible early loans. In
this scenario, the contact between Indo-European and
Caucasian languages could be accounted for by assuming that
there was a northward migration of Caucasian-speaking people
into Indo-European-speaking territory, somewhat along the
lines that Colarusso has suggested. In my opinion, scenario
(C) is less attractive than scenario (A).

The above map shows the area to which Indo-European languages had
spread by the first century BCE (this is taken from Villar 1991 :17).

CONCLUSION

Thus, Proto-Indo-European is not to be seen as a
Caucasian language but rather as a caucasianized form of the
Indo-Uralic (more accurately, Pre-Indo-European) dialect that
had spread to the shores of the Black Sea. On the basis of
grammatical, pronominal, and lexical cognates with other
Eurasiatic languages, Proto-Indo-European is indeed to be
classified as a Eurasiatic language, exactly as Greenberg has
done. However, due to long-term contact with and extensive
borrowing from Caucasian languages, Proto-Indo-European
developed many characteristics which set it apart from the
other branches of Eurasiatic and gave it a Caucasian- like
appearance. In other words, we are talking about an early
Pontic-Caucasus Sprachbund area which included languages
belonging to more than one language family. At the present
time, the same phenomenon may be observed in the case of
Ossetic, which shares many areal features with surrounding
languages, though it is an Iranian language and not genetically
related to its immediate neighbors (cf. Comrie 1981:158-179

MOTHER TONGUE

Issue 22, May 1994

for details).

Now, in view of the additional cognates listed in this
paper, it may be possible that Proto-Northwest Caucasian
belongs with Nostratic, if not as a daughter branch, possibly as
a sister — it all depends upon the level at which one wishes to
reconstruct a proto-language, upon how one sets up subgroups,
and upon how one establishes the special relationships among
members of the various subgroups. If this should turn out to be
the case, then Proto-Indo-European and Proto-Northwest
Caucasian would indeed be genetically related as descendants
from a common ancestor, without necessarily implying a closer
relationship. On the other hand, the lexical parallels in which
cognates can be found in other Nostratic languages may simply
be due to borrowing by Northwest Caucasian, which would
mean that the case for genetic relationship would be
substantially weakened.

NOTES

*I would like to thank Merritt Ruhlen for commenting
on earlier drafts of this paper. His advice was extremely
valuable and saved me from making a number of foolish
mistakes. I alone am responsible for any errors that may exist
in this paper. I would also like to thank Johanna Nichols for
sending me a copy of the manuscript of her unpublished paper
entitled "The Origin and Dispersal of Indo-European".

1) Here, I am using the term "blend" to conform with
Colarusso — nowadays, the term "convergence" would be used
to describe this kind of language contact.

2) The term "tectal" is being used as a replacement for
the term "guttural", in accordance with current usage in Indo-
European circles (cf, for instance, Lehmann 1993:19-20, 57,
and 100-101).

3) Merritt Ruhlen has informed me (personal com-
munication) that Starostin follows Illic-Svityc in positing a
three-way contrast in obstruents for Proto-Altaic. As we were
going to press, Ruhlen sent me a photocopy of Starostin' s book
— I deeply appreciate his kindness.

REFERENCES

Bomhard, Allan R. 1988. "The Prehistoric Development of
the Athematic Verbal Endings in Proto-Indo-European".
A Linguistic Happening in Memory of Benjamin
Schwartz, edited by Yoel L. Arbeitman, pp. 475-488.
Louvain: Publications Linguistique de Louvain.

Bomhard, Allan R. and John C. Kerns. 1994. The Nostratic
Macrofamily: A Study in Distant Linguistic Relationship.
Berlin, Amsterdam, New York, NY: Mouton de Gruyter.

Buck, Carl Darling. 1949. A Dictionary of Selected Synonyms
in the Principal Indo-European Languages. Chicago, IL:
University of Chicago Press.

Colarusso, John. 1975. The Northwest Caucasian Languages:
A Phonological Survey. Ph.D. dissertation, Harvard
University.

Colarusso, John. 1994. "Phyletic Links between Proto-Indo-
European and Proto-Northwest Caucasian". Mother
Tongue 21:8-20.

Collinder, Bjorn. 1934. Indo-Uralisches Sprachgut Die
Urverwandtschaft zwischen der indoeuropaischen und
der uralischen (finnisch-ugrisch-samoyedischen) Sprach-
familie. Uppsala: Uppsala Universitets Arskrift.

Comrie, Bernard (ed.) 1981. The Languages of the Soviet
Union. Cambridge: Cambridge University Press.

Decsy, Gyula. 1990. The Uralic Protolanguage: A
Comprehensive Reconstruction. Bloomington, IN:
Eurolingua.

Dolgopolsky, Aaron. 1984. "On Personal Pronouns in the
Nostratic Languages". Linguistica et Philologica.
Gedenkschrift fur Bjorn Collinder (1894-1983), edited by
Otto Gschwantler, Karoly Redei, and Hermann Reichert,
pp. 65-112. Vienna: Wilhelm Braumuller.

Gamkrelidze, Thomas V. and Vjaceslav V. Ivanov. 1984.
Indo-evropejskij jazyk i indoevropejcy: Rekonstrukcija i
istoriko-tipologiceskij analiz prajazyka i protokul 'tury
(Indo-European and Indo-Europeans: A Reconstruction
and Historical Typological Analysis of a Protolanguage
and a Proto-Culture) . 2 vols. Tbilisi: Publishing House
of the Tbilisi State University.

Greenberg, Joseph H. Forthcoming: Indo-European and Its
Closest Relatives: The Eurasiatic Language Family.
Stanford, CA: Stanford University Press.

Hajdu, Peter. 1975. Finno-Ugrian Languages and Peoples.
Translated and adapted by G. F. Cushing. London:
Andre Deutsch Ltd.

Illic-Svityc, V. M. 1971- . Opyt sravnenija nostraticeskix
jazykov (semitoxamitskij, kartvel'skij, indoevropejskij,
uraVskij, dravidskij, altajskij) (An Attempt at a Com-
parison of of the Nostratic Languages [Hamito-Semitic,
Kartvelian, Indo-European, Uralic, Dravidian, Altaic]).
3 vols. Moscow: Nauka.

Joki, Aulis. 1973. Uralier und Indogermanen. Die alteren
Beriihrungen zwischen den uralischen und indoger-
manischen Sprachen. Helskinki: Suomalais-Ugrilainen
Seura.

Kerns, John C. 1985. Proto-Indo-European Prehistory.
Cambridge: Heffer and Sons.

Kuipers, A[ert] H. 1975. A Dictionary of Proto-Circassian
Roots. Lisse: Peter de Ridder Press.

Lehmann, Winfred P. 1993. Theoretical Bases of Indo-
European Linguistics. London and New York, NY:
Routledge.

Mann, Stuart E. 1984-87. An Indo-European Comparative
Dictionary. Hamburg: Helmut Buske Verlag.

Nichols, Johanna. 1993. "The Origin and Dispersal of Indo-
European." Manuscript.

Pokorny, Julius. 1959. Indogermanisches etymologisches
Worterbuch. Vol. I. Bern and Munich: Francke Verlag.

Ruhlen, Merritt. 1987. A Guide to the World's Languages.
Stanford, CA: Stanford University Press.

Szemerenyi, Oswald. 1990. Einfuhrung in die vergleichende
Sprachwissenschaft. 4th edition. Darmstadt: Wissen-
schaftliche Buchgesellschaft.

Villar, Francisco. 1991. Los indoeuropeos y los origenes de
Europa. Lenguaje e historia. Madrid: Gredos.

MOTHER TONGUE

Issue 22, May 1994

IS PROTO-INDO-EUROPEAN

RELATED TO PROTO-NORTHWEST

CAUCASIAN?

MERRITT RUHLEN
Palo Alto, CA

In a recent article, John Colarusso proposes that Proto-
Indo-European (henceforth PIE) and Proto-Northwest
Caucasian (henceforth PNWC) were descendants of a Proto-
Pontic language that was spoken north of the Black Sea around
10,000 years ago (Colarusso 1992). His conclusion is that "PIE
and PNWC are genetically related at a phyletic level." I
personally have no doubts that PIE and PNWC are genetically
related, and have shown as much in a joint article with John
Bengtson (Bengtson and Ruhlen 1994). But I also have no
doubt that Colarusso's mythical Proto-Pontic never existed —
north of the Black Sea or anywhere else.

What Colarusso fails to realize is that these two
families are related to one another without in any way
constituting a valid linguistic taxon, or family. They are related
to each other because they are members of larger families,
which are themselves in turn related to one another. They are
not related to each other because each of them is genetically
closer to the other than either is to any other family of
languages — and indeed such is not the case. It is worth noting
that Sergei Starostin (1989) has also shown that PIE is related
to PNWC, but his line of evidence is very similar to that of
Bengtson and Ruhlen (1994) in that what he shows is that
Nostratic is related to Sino-Caucasian — the first including
PIE, the second encompassing PNWC — not that there is any
direct connection between PIE and PNWC.

Colarusso's direct approach to distant genetic
relationships is really a denial of the evolutionary process itself,
for relationships, if they are anything but transparent, are not of
the immediate nature that Colarusso envisages. No biologist
doubts that pigs and humans are related, but it is not because
pigs and humans derive from Proto-Porco-Human, but rather
because pigs and humans belong to groups, which in turn
belong to groups, which ultimately derive from a single group.
No biologist would ever even think of the idea of comparing
pigs and humans by themselves, much less actually do it.
Further discussion of the nature of linguistic taxonomy — and
its relationship to "traditional" historical linguistics — may be
found in Ruhlen (1994a, 1994b).

This is one of the fundamental differences between
biological taxonomists and historical linguists of the Eric Hamp
school. It is certainly not surprising that Hamp appears to have
inspired Colarusso's paper, for it is reminiscent of the putative
Mayan-Chipaya, and Mayan-Mapudungu, hypotheses in the
Americas, both of which were supported by Hamp (1967,
1971). According to the first, the Mayan family was alleged to
form a family with the Chipaya language of Bolivia (Olson
1964, 1965), while the latter hypothesis (Stark 1970) connected
Mayan with Mapudungu. When one approaches these
problems more soberly (Greenberg 1987), it turns out that
Mayan, Chipaya, and Mapudungu are each members of

different branches of Amerind (Penutian, Equatorial, and
Andean, respectively) and what they share is their common
Amerind inheritance, not the special relationship envisaged by
Hamp. These cases are directly parallel to Colarusso's Proto-
Pontic and illustrate the absurdity of arbitrary binary
comparisons.

What is most disturbing about Colarusso's article is
that he operates as if nothing were known about the
classification of the world's languages. His references cite none
of the standard works on linguistic taxonomy (e.g. Greenberg
1957, 1963, Voegelin and Voegelin 1977, Ruhlen 1987), nor
any of the abundant literature on the closest relatives of the
Indo-Europeans, or of the Northwest Caucasians. In fact, he
dismisses this literature in a footnote wherein he refers to "the
folly of such grand lumping schemes" as Nostratic (or
Eurasiatic) and Dene-Caucasian, without citing any of the
relevant literature on these two families.

But one does not have to leap to the limits of Dene-
Caucasian to see the folly of Colarusso's approach. It is now
generally agreed that Northwest Caucasian languages are most
closely related to Northeast Caucasian languages as the
Caucasian family (Comrie 1981:198, Catford 1991:234). The
earlier work of Trubetskoy (1930) and others had long ago
established this fact, but the more refined and complete study
of Nikolaev and Starostin (1992) demonstrates the relationship
beyond any shadow of a doubt, and this fact by itself 'shows that
Colarusso's hypothesis is incorrect in that PNWC is
transparently closer to Proto-Northeast Caucasian (henceforth
PNEC) than it is to PIE.

On the other side of the equation, the case is equally
clear, though a good deal more controversial. Despite the fact
that linguistic similarities among Eurasian families — many of
which had already been noticed in the nineteenth century — led
Holger Pedersen to posit, early in the present century, a
Nostratic family that included Indo-European; and despite the
thorough refinement and advancement of this hypothesis by
Vladislav M. Illic-Svityc (1965, 1971-1984) and Aaron B.

Dolgopolsky (1964, 1984), as well as the complementary work
of Allan Bomhard (1991) on the same theme; and despite the
similar results obtained by Joseph Greenberg using multilateral
comparison to identify a Eurasiatic language family that is very
similar to Nostratic (Greenberg, to appear), Indo-Europeanists,
and Colarusso, act as if none of this evidence even exists. One
might call it the Clinton model of historical linguistics, "Don't
ask, don't tell, and lie about it if they find out." But this ostrich
approach to historical linguistics cannot do away with — much
less explain — the fact that the Indo-European pronominal
pattern m-/t- 'I, you' is shared, in whole or in part, by numerous
other Eurasian language families (e.g., Kartvelian, Uralic,
Altaic, Korean, Japanese, Gilyak, Chukchi-Kamchatkan, and
Eskimo- Aleut). Nor can it hide the fact that dozens of other
grammatical formatives, and hundreds of other lexical items,
are shared by these same families, and are absent from the
world's other language families. Nor can it hide the fact that
PNWC *sA T resembles such Northeast Caucasian forms as
Proto-Nakh *so, Proto-Lezghian *zo-n, and Xinalug zi far
more than it does Eurasiatic *mi. Nor does it conceal the fact
that PNWC *wA 'you' resembles Lak wi- 'you,' Proto-

MOTHER TONGUE

Issue 22, May 1994

Lezghian *uo-n 'you,' Xinalug wi 'you', and probably Proto-
Nakh *waj 'we (inc.)' (if this form derives from a combination
of 'you' + 'me,' a typologically plausible origin of inclusive
pronouns) far more than it does Eurasiatic *ti. Nor can it
obscure the fact that both of these Proto-Caucasian pronouns
are attested in other branches of Dene-Caucasian. The first-
person pronoun is also found in Burushaski fa, Sino-Tibetan
(e.g. Newari fi), Proto-Yeniseian *?ad z , and Na-Dene (e.g.
Proto-Athabaskan *sl.) The second-person pronoun is seen in
Burushaski u-n, Proto-Yeniseian *?u, and Na-Dene (Tlingit

wae) .

Indeed, any objective taxonomist who compares
Colarusso's feeble comparisons between PIE and PNWC with
the robust Caucasian comparative dictionary of Nikolaev and
Starostin (1992), and at further remove, the Dene-Caucasian
comparisons of Starostin (1982, 1984), Nikolaev (1986),
Bengtson (1991a, b), Shafer (1952, 1957), and Cirikba (1985)
on one hand, and the Nostratic/Eurasiatic comparisons, on the
other, will find that Colarusso's comparisons are few and
unconvincing when compared with the aforementioned body of
evidence. Rather than using basic vocabulary, such as personal
pronouns, body parts, and natural phenomena, Colarusso uses
meanings such as "giant, cattle," and "(hard) metal" in his 20
"conventional cognates." Furthermore, the phonological
connections are often less than transparent. Thus we are asked
to believe that PIE *dwo- 'two' derives from earlier stages:
*t'o? w < *f? w s < *t'q'o, this latter form conveniently being

identical with Proto-Pontic (and PNWC) *t'q 'o 'two.' If we are
to believe all this, then PIE has undergone a whole series of
radical sound changes during the same time that PNWC
remained unchanged.

In sum, both the semantics and phonetics of
Colarusso's comparisons are largely unconvincing. Some of the
correct comparisons are in fact much more widespread than
simply these two families and thus provide no evidence for
Proto-Pontic. One such example is Colarusso's Proto-Pontic
*pa 'son, child.' This same root was posited for Nostratic by
Illic-Svityc (1965) and for Sino-Caucasian by Starostin (1984),

who also noted the similarities to the Nostratic forms in
Starostin (1989). Greenberg (1987:199-200) showed that the
same root is widespread in the Amerind family, and I have
compared the Amerind material with the Eurasiatic forms
(Ruhlen 1994c). In Amerind this root has undergone two
interesting developments. First, the meaning has shifted from
'son, child' to 'brother'; and second, reduplication is used with
the root to distinguish 'younger brother' from 'older brother,' as
in Proto-Oto-Manguean *po 'younger brother' vs. *papi 'older
brother.' In addition to these objections, Colarusso's Pre-PIE
reconstructions are tortured and convoluted — as if their sole
purpose was to make them seem similar to PNWC, perhaps not
surprising in a paper that the author admits was initially
inspired by typological similarities. It is a pity that in the late
twentieth century such absurd proposals as Proto-Pontic are not
only advanced by supposed experts, but are even discussed
seriously by other supposed experts.

REFERENCES

Bengtson, John D. 1991a. "Notes on Sino-Caucasian," in:
Vitaly Shevoroshkin, (ed.), 1991: 67-129.

Bengtson, John D. 1991b. "Some Macro-Caucasian
Etymologies," in Vitaly Shevoroshkin, ed., Dene-Sino-
Caucasian Languages, pp. 130-41. Bochum:
Brockmeyer.

Bengtson, John D., and Merritt Ruhlen. 1994. "Global
Etymologies," in: Ruhlen (1994a), pp. 277-336.

Bomhard, Allan R. 1991. "Lexical Parallels between Proto-
Indo-European and Other Languages," in: L. Isebaert
(ed.), Studia Etymologica Indoeuropaea, pp. 47-106.
Leuven: Peeters.

Catford, J. C. 1991. "The Classification of Caucasian
Languages," in: Sydney M. Lamb and E. Douglas
Mitchell (eds.), Sprung from Some Common Source:
Investigations into the Prehistory of Languages, pp. 232-
68. Stanford, CA: Stanford University Press.

Cirikba, V. A. 1985. "Baskskij i severokavkazskie jazyki,"
Drevnjaja Anatolija, pp. 95-105. Moscow.

Colarusso, John. 1992. "Phyletic Links between Proto-Indo-
European and Proto-Northwest Caucasian," in: Howard I.
Aronson (ed.), The Non-Slavic Languages of the USSR:
Linguistic Studies (Second Series), pp. 19-54. Chicago,
IL: Chicago Linguistic Society. Reprinted in Mother
Tongue 21 :8-20.

Comrie, Bernard. 1981. The Languages of the Soviet Union.
Cambridge: Cambridge University Press.

Dolgopolsky, Aaron. 1964. "Gipoteza drevnejsego rodstva
jazykovyx semei severnoj Eurasii s verojatnostnoj tocki
zrenija," Voprosy Jazykoznanija 2:53-63. [English
translation in: Vitalij V. Shevoroshkin and T. L. Markey,
(eds.), 1986, Typology, Relationship and Time, pp. 27-50.
Ann Arbor, MI: Karoma.]

Dolgopolsky, Aaron. 1984. "On Personal Pronouns in the
Nostratic Languages," in: Otto Gschwantler, Karoly
Redei, and Hermann Reichert, (eds.), Linguistica et
Philologica, pp. 65-1 12. Vienna: Braumiiller.

Greenberg, Joseph H. 1957. "Genetic Relationship among
Languages," in Joseph H. Greenberg, Essays in
Linguistics, pp. 35-45. Chicago, IL: University of
Chicago Press.

Greenberg, Joseph H. 1963. The Languages of Africa.
Bloomington, IN: Indiana University Press.

Greenberg, Joseph H. 1987. Language in the Americas.
Stanford, CA: Stanford University Press.

Greenberg, Joseph H. To appear. Indo-European and Its
Closest Relatives: The Eurasiatic Language Family.
Stanford, CA: Stanford University Press.

Hamp, Eric P. 1967. "On Maya-Chipayan," International
Journal of American Linguistics 33:74-16.

Hamp, Eric P. 1971. "On Mayan- Araucanian Comparative
Phonology," International Journal of American
Linguistics 37:156-59.

Illic-Svityc, Vladislav M. 1965 [1967]. "Materialy k
sravnitel'nomu slovarju nostraticeskix jazykov,"
Etimologija (Moscow) 1965:321-96. [English translation

MOTHER TONGUE

Issue 22, May 1994

in: Vitaly Shevoroshkin, (ed.), 1989, Reconstructing

Languages and Cultures, pp. 125-76. Bochum:

Brockmeyer.]
Illic-Svityc, Vladislav M. 1971-1984. Opyt sravnenija

nostraticeskix jazykov, 3 volumes. Moscow. [English

translation of the reconstructions, and a semantic index to

them, in: Shevoroshkin, (ed.), 1990, Proto-Languages

and Proto-Cultures, pp. 138-67. Bochum: Brockmeyer.]
Nikolaev, Sergei L. 1986. "Sino-kavkazkie jazyki v Amerike,"

ms. [English translation in: Vitaly Shevoroshkin, (ed.),

1991, Dene-Sino-Caucasian Languages, pp. 42-66.

Bochum: Brockmeyer.]
Nikolaev, Sergei L., and Sergei A. Starostin. 1992. A

Caucasian Etymological Dictionary. Manuscript.
Olson, Ronald D. 1964. "Mayan Affinities with Chipaya of

Bolivia: Correspondences," International Journal of

American Linguistics 30:313-24.
Olson, Ronald D. 1965. "Mayan Affinities with Chipaya of

Bolivia: Cognates," International Journal of American

Linguistics 31:29-38.
Ruhlen, Merritt. 1987. A Guide to the World's Languages,

Volume 1: Classification. Stanford, CA: Stanford

University Press.
Ruhlen, Merritt. 1 994a. On the Origin of Languages: Studies in

Linguistic Taxonomy. Stanford, CA: Stanford University

Press.
Ruhlen, Merritt. 1994b. The Origin of Language. New York,

NY: John Wiley & Sons.
Ruhlen, Merritt. 1994c. "The Linguistic Origins of Native

Americans," in Ruhlen 1994a, pp. 207-41.
Shafer, Robert. 1952. "Athabascan and Sino-Tibetan,"

International Journal of American Linguistics 18:12-19.
Shafer, Robert. 1957. "Note on Athabascan and Sino-Tibetan, "

International Journal of American Linguistics 23:1 16-17.
Stark, Louisa R. 1970. "Mayan Affinities with Araucanian," in:

Papers from the Sixth Regional Meeting of the Chicago

Linguistic Society, pp. 57-69. Chicago, IL: Chicago

Linguistic Society.
Starostin, Sergei A. 1982. "Prajenisejskaja rekonstruktsija i

vnesnie svjazi enisejskix jazykov," Ketskij sbornik.

Leningrad, pp. 144-237.
Starostin, Sergei A. 1984. "Gipoteza o geneticeskix svjazjax

sinotibetskix jazykov s enisejskimi i severnokavkazskimi

jazykami," Lingvisticeskaja rekonstruktsija i drevnejsaja

istorija vostoka 4. Moscow, pp. 19-38. [English
translation in: Vitaly Shevoroshkin, (ed.), 1991, Dene-
Sino-Caucasian Languages, pp. 12-41. Bochum:
Brockmeyer.]

Starostin, Sergei A. 1989. "Nostratic and Sino-Caucasian," in:
Vitaly Shevoroshkin, (ed.,), Explorations in Language
Macrofamilies, pp. 42-66. Bochum: Brockmeyer.

Trubetskoy, Nikolai S. 1930. "Nordkaukasische
Wortgleichungen," Wiener Zeitschrift fur die Kunde des
Morgenlandes 37.

Voegelin, C. F., and F. M. Voegelin. 1977. Classification and
Index of the World's Languages. New York, NY:
Elsevier.

COMMENT ON COLARUSSO 1994

JOHND.BENGTSON
Minneapolis, MN

Editor Allan Bomhard has invited me, as a proponent
of the Dene-Caucasian hypothesis, to comment on John
Colarusso's hypothesis of a genetic link between Proto-Indo-
European and Proto-Northwest Caucasian (hereafter simply
West Caucasian or WCauc), which was reprinted in the last
issue of Mother Tongue.

The hypothesis of genetic links between I(ndo-)
E(uropean) and Caucasic languages dates long before 1964. C.
C. Uhlenbeck (1937) thought that IE consisted of two distinct
elements, A (which he identified with "Ural-Altaic", i.e.,
Nostratic) and B (identified with Caucasic languages in
general, in the same sense Colarusso uses). To Uhlenbeck,
Indo-European was presumably a Creole language, or series of
Creole languages, that developed from the interaction of the A
invaders with the indigenous B populations all over Europe.
By the way, Uhlenbeck also connected the Eskimo language(s)
with element A (Nostratic), and Basque with element B
(Caucasic).

Anybody who is tempted by Colarusso's hypothesis
should look at an article published by Prince Nikolai
Trubetzkoy over six decades ago (Trubetzkoy 1930). This is
the work that firmly established, in my opinion, the genetic link
between West Caucasian and East Caucasian, on the basis of
one hundred lexical comparisons. As in any pioneering study,
there were some errors and incorrect comparisons, but we now
have the good fortune of a recent edition (Trubetzkoy 1987)
with corrections and additions by Sergei Starostin ((1987),
making use of the new reconstructions of both families.
Because these works are inaccessible to many readers of
Mother Tongue, I will list some of Trubetzkoy 's comparisons
here. (Five comparisons, §§2, 6, 14, 17, 21, have been added.
Basque, Burushaski, and other Dene-Caucasian forms are also
cited, where relevant):

1. "I" (1st person singular): WCauc *sV (Circassian] sa) =
ECauc *zo (Chechen so, Khinalug za, etc.); cf. Burushaski
ja, je, (W[erchikwar]) za; Yen(iseian) *?aj; N(a)-D(ene):

Athabaskan *sL

2. "two": WCauc *(t)q w V (Circ t'™d) = ECauc *q'w§ (Lak
k'i, Archi q' w e); cf. Basque bi (from *G w i, assimilated to

bat "one", like English four, five, Latin quattuor, quinque);
S(ino-)T(ibetan) *k-niy(s); Yen *xi-na; ND: Athabaskan
*qi.

3. "thou" (2nd person singular): WCauc *wV (Circ wd) =
ECauc *uo- (Archi u-n, Khinalug vd); cf. Burushaski w-n;
Yen *?aw/*?u; ND: Tlingit wa-e.

4. "you" (2nd person plural): WCauc *s w 9 (Circ s w a) =
ECauc *zwV (Lak zu) - Basque zu; cf. Sumerian za-, ze-,
ze- "thou", -zu "thy".

5. "what?": WCauc *sV(ybybhsa) = ECauc *sV(Dargise);
cf. Basque ze-r; Yen *?as-/*sV- (interrogative); ND:

MOTHER TONGUE

Issue 22, May 1994

Tlingit daa-sa "what", etc.

6. (interrogative): WCauc *-an(3)- "when?" = ECauc *nV
(Lezgi ni "who?"); cf. Basque no-r "who?", no-iz
"when?"; Burushaski (W) ana "where, whither?";
Sumerian a-na, en, en-na "what"; ST: Tibetan na
"when?"; Yen *?an- "who".

7. "tongue": WCauc *bdzV (Circ bzd) = ECauc *melc'i
(Andi mic'i, Tabasaran/ne/j); cf. Basque mihi; Burushaski
melc "jaw"; (?) ST *mlay "tongue".

8. "name": WCauc *(p-)c 'a (Ubykh/? 'c 'a) = ECauc *3wer?i
(Andi e'er); cf. Burushaski -car "voice".

9. "eye"; WCauc *b-l'a (Abaza la) = ECauc *?wil?i (Dargi
huli, Chechen bfarg); cf. Basque begi (from *beryi);
Burushaski -lei, -i'U.

10. "heart": WCauc *g w d (Circg^a) = ECauc *jerk'wi (Lezgi

rik', Chechen dog); (?) Basque bi-rika "lung".

11. "tooth": WCauc *ca (Circ ea) = ECauc *cim V (Andi sol,

Chechen cerig); Basque hortz [horc] (regular
correspondence to metathesized Cauc *Sf/cV);

12. "not": WCauc *md- (= Circ md-) = ECauc *mV (Chechen
ma); cf. Burushaski be.

13. "louse": WCauc *c'a = ECauc *nemjV (Lak nac\

Chechen mezi); cf. Basque bartz "nit" (from *welje >
*wen3e > *men3e > ^nemje, by a series of assimilations
and dissimilations; further Khinalug nimc' ~ lime'
"louse"); the original shape *welje or *welc'e is
confirmed by ND: Tlingit wes ' "louse".

14. "die": WCauc *if 9 V (Circ K'd-n "to die, to kill") = ECauc
*?i(w)X'V (Hunzib iX'a); cf. Basque hil; ST: Old Chinese
*Kiy "corpse"; ND: Haida K'a-daa "to kill" (with plural

object).

15. "moon": WCauc *maza (Circ ma'za) = ECauc *w3inc'o
(Avar moc 9 , Lezgi varz); (?) Old Chinese *y w at.

16. "hand": WCauc *q'a ~ *qa = ECauc *q'wel_V; cf. Basque
u-khare, etc. (cf. §22, below).

17. "night": WCauc: Abkhaz -Xa = ECauc (Lezgian) *Xam;

cf. Basque gau, gab-.

18. "horn": WCauc *q™a (Ubykh qa) = ECauc *qwa(hV) (Lak
#/, Dargi #e); cf. Yen *Xo?.

19. "full": WCauc *zK (Ubykh -za) = ECauc *?oc' V (Lak

b-uwe'u "full", Chechen d-uza "to fill"); cf. Basque oso
"whole, entire, healthy"; Yen *?ute "full"; ND: (?) Haida

st 'ah "full", Sarsi -c'ist.

20. "sun": WCauc *r9ya (Ubykh dya) = ECauc *(wi-)raoV

(Lezgi ray, Archi barq); cf. Basque argz "light", argi-zagi,
hil-argi "moon".

21. "ear": WCauc *1'V (Abkhaz -temha) = ECauc *lerhV

(Dargi lihi, Chechen lerg); cf. Basque be-larri.

Note that these comparisons all fall within Dolgopolsky's
hierarchy of the twenty-five historically most stable meanings

(Dolgopolsky 1964, 1986; Shevoroshkin and Markey
1986:xvii-xviii), the semantic fields that are the most likely to
retain the genetic core of a language. The twenty-one
comparisons listed here demonstrate the unity of WCauc with
ECauc, Basque, and other languages in a macrofamily or
macrophylum called Dene-Caucasian (Starostin 1982, 1984,
1989; Nikolaev 1991; Bengtson 1990, 1991a, etc.). One
searches in vain for anything resembling the familiar Indo-
European "tongue" (*dnghu), "eye" (*oq~-), or "tooth"

{*edont-)\ (The Indo-European reconstructions are from A.
Walde, Vergleichendes Worterbuch der indogermanischen
Sprachen.)

These comparisons stand in strong contrast to those of
Colarusso, which include such meanings as "house" (§68),
"giant" (§70), "cattle" (§73), "metal" (§§77, 78), all highly
susceptible to cultural and areal diffusion. Indeed, of
Colarusso's twenty "conventional cognates" (§§64-83), only
one (§75 "two") falls within Dolgopolsky's hierarchy. His
"morphological cognates" do include some personal pronouns
(§33), but I find the comparisons tortured and implausible,
compared with the straightforward parallels involving WCauc,
ECauc, and other DC languages (see my §§1, 3, 4 above).

What about Colarusso's list of "morphological
cognates", which Bomhard, in his comments on Colarusso's
paper, finds convincing? Some of these "cognates" are of a
general and typological nature (e.g., the opposition of thematic
and athematic nouns and verbs) that could be found in any
number of languages without being evidence of genetic
relationship. Others are phonologically implausible, for
example (as Bomhard admits), the comparison of IE *s- with
WCauc *(-y)-K h - (§52). Yet others may indeed be historically

valid, but not, I think, indicative of a close relationship between
IE and WCauc Some of these are reminiscent of comparisons
made by Robert Shafer (1963, 1965) between IE and Sino-
Tibetan. For example, the accusative -m (Colarusso §31) is
found in the ST languages Mishing and Nyising (Shafer
1965:447, §2); and the IE sigmatic aorist (Colarusso §44) also
has ST parallels (Shafer 1965:451, §16). Should "Pontic" or
Nostratic then be extended to include ST as well?

I think a better solution is that of Starostin (1989),
proposing a deep relationship between Nostratic (including IE)
and Dene-Caucasian (including Cauc and ST), a proposal that
is in fact a refinement of Shafer's "Eurasial". Starostin's
comparisons encompass several of the parallels adduced by
Colarusso (and Bomhard), for example, the negative/
prohibitive particle *mV- (Starostin 1989:64, §2) and the
numeral "two" (Starostin 1989:62, §180). While I reject
Colarusso's claim of a close genetic tie between IE and WCauc,
I do not consider his efforts a total loss by any means. I think
they can be useful for (a) typological and areal connections
between IE and WCauc, and (b) evidence of a deeper genetic
tie between Nostratic and Dene-Caucasian.

This problem is instructive, because it brings us back
to the basic principles of genetic classification. Classification
is not just a matter of piling up masses of etymologies and
continually adding more languages to one's chosen family or
macrofamily. The logical extension of Colarusso's results
would produce a virtually endless string of languages, but no

MOTHER TONGUE

Issue 22, May 1994

clue as to the correct subgrouping. (The negative mV is found
in Amerind as well. Do we then add Amerind to Nostratic?)
The problem is to evaluate the parallels according to a
hierarchy of stability such as Dolgopolsky's, or the Swadesh
lists: "tongue", "name", "eye" are more indicative of genetic
relationship than "house", "cattle", "metal"! When only words
within the hierarchy are compared, it becomes quite clear that
West Caucasian belongs to a group with East Caucasian (and,
in my view, Basque and the other DC languages), while Indo-
European does not.

Another principle involves structural compatibility.
Edward Sapir (e.g., 1913, 1915) always insisted on establishing
a structural pattern as the first step in the demonstration of
genetic affinity. For example, in Dene-Caucasian there is
abundant evidence of class markers which were prefixed to
nouns, and also appear in some verb forms (Toporov 1971;
Bengtson 1991a, 199 Id, 1991e, 1993). This system is very
much alive in many Cauc languages, in Burushaski, and in
Yeniseian. In Basque (noun prefixes a-, e-/i- 9 be-/bi-, o-/u-)
and in Sino-Tibetan (noun prefixes a-, b-, r-, etc.), these
elements appear as "fossilized" and lexicalized remnants. This
type of structure is alien to the early stages of Indo-European
and other Nostratic languages.

Colarusso has not demonstrated that IE belongs in a
family with WCauc (except, perhaps, at a remote level),
because he has not shown that IE is closer to WCauc than it is
to other families. We must not be fooled by typological
similarities (remember the problems we had in Africa and
Southeast Asia), nor by similarities that are too universal to be
diagnostic (e.g., Bomhard's *k'ak'a "to chirp", *hana
"mother", etc.). There have to be some agreements in basic
vocabulary and grammatical structure.

As a coda, here are some Dene-Caucasian
comparisons involving West Caucasian material. Several of
them are drawn from "Lexica Dene-Caucasica" by Vaclav
Blazek and myself (Blazek and Bengtson 1995), which we

think further solidifies the comparative lexicon of the Dene-
Caucasian macrophylum. The reader can judge whether these
are more or less convincing than the WCauc-IE comparisons
offered by Colarusso (and Bomhard):

22. WCauc *q'a ~ *qa "hand" (Circ ?a, Ubykh q'a-p'a) I
ECauc *q'wel_V "hand" (Avar q'wal, etc.) / Basque u-
kharai, u-khare, u-kari (combinatory form u-k[h]al-)
"wrist" / Yen(iseian) *xire "hand": Ket ¥l'(i) 9 Arin kara-,
koro- (in compounds) / (?) S(ino-)Tibetan: Tibetan khyor
"palm, handful" (Starostin 1982:209).

23. WCauc: Circ *s w td (Adyghe § w t9, Kabardian ftd

"genitals") / Basque iztai "anus", izter "thigh" / Burushaski
-asc-irj, (W) -ast-ig "small of the back, loins, reins, waist"

/ ST: Tibetan sta-zur "hip, hipbone" (entire comparison by
Bouda 1950, §204 and 1954, §19).

24. WCauc: Abkhaz -k'dk'a- "female breast" / ECauc
*kik[u]: Hunzib kigla id., Lak kuku "nipple" / Basque

kholko, golko (from *khoklo ?) "breast, bosom" / ND:
Haida k'wk "heart", Tlingit kegu "lungs" (Cirikba

1985:99, §44; Blazek and Bengtson 1995, §23).

25. WCauc *c'ac'a "kidney" (Abkhaz a-c'ac'a, Kabardian za
zaj) I ECauc *c 9 ec'V (Dargi ur-cec, Khinalug c'ic'in
"kidney") / Yen *sisa(l)- "lights" (lungs) (Kott sica'tn,
Arin sisali) I ND: Haida cay "kidney"; Athabascan: Sarsi
c'uza id. (Starostin 1982:224; Blazek and Bengtson 1995,
§28).

26. WCauc: Circ *c'a§ w a (Adyghe c'd§ w a, Kabardian §'afa

"human skin, body") / Yen *sa's "(animal) skin": Ket
sa'si, Kott set I ND: Tlingit jas "skin"; Eyak -sic' "fish
skin"; Athabaskan: Galice -sas "(animal) skin", Hupa sic 9
"skin, bark" (Blazek and Bengtson 1995, §36).

27. WCauc *Ia "foot" (Ubykh ia) I ECauc *MhV (Avar lar-k'
"sole", Chechen lar "trace") / Basque *lor- in lorr-ats
(Vizcayan, Guipuzcoan) "track, trail, trace, scent" / ST
*laH "foot" (Starostin): Chepang la, Dimasa yd) I ND:
Eyak/3? "thigh" (Starostin 1984; 1991, §19; Blazek and

Bengtson 1995, §41).

28. WCauc: Circ *k h ax w (Kabardian ita^ "dry twigs") / N(a-)

D(ene): Haida kuk "firewood"; Eyak kuk-} "dry wood,
firewood", Navajo clz, etc. (Blazek and Bengtson 1995,

§71).

29. WCauc *k'ank'a "egg" / ECauc *k'erk'enV (Andi
k'ork'on "egg", Avar k'ork'onu "grape, berry") / Basque
(Vizcayan) kankano "large fruitstone, kernel, almond" /
Burushaski kaka-yo "(walnut) kernel" / ND: Haida
k'a'nk'a-y ~ k'a-nk'a-n "unripe berries"; Athabaskan:
Navajo -k 9 g? "seed, pit" (Blazek and Bengtson 1995,

§74).

30. WCauc: Circ *q'a§' w 3 "sweet" / ECauc *q'(w)V16'6'V
"sour" (Lak q'urc'-) I Basque gozo ~ goxo [goso] "sweet,

tasty" / Burushaski gas-ar-um "salt-sweet" / Sumerian
kus 6 "honey, sweet" /ND: Eyak q'ihj "rancid, bitter, sour,
spoiled"; Athabaskan: Navajo k 9 qs, -k'gz "sour, salty,
brackish", Mattole -k'o?3 "sweet" (Nikolaev 1991:57,
§11.12; Blazek and Bengtson 1995, §187).

31. WCauc: Abkhaz -k w a (plural of "non-reasonable" class) /
Basque -k (noun plural suffix) / Burushaski -ko(rj) (plural
of certain nouns) / ND: Tlingit -x\ -x' w (plural or
collective suffix); Athabaskan: Navajo -ke, -kei (noun
plural suffix) (Dumezil 1933:135; Blazek and Bengtson

1995, §212).

32. WCauc: Abkhaz -as, -s (instrumental suffix) / ECauc:
Chechen -s (animate ergative), Hurrian -(u)s (ergative) /
Basque -z, -ez (instrumental) / ST: Tibetan -s
(instrumental), Dhimal -so id., Kanauri -s (agent) / Yen:
Ket -as (instrumental-comitative), Kott -os (comitative), -
s(e) (instrumental) (Tailleur 1958:418; Cirikba 1985:95;
Blazek and Bengtson 1995, §213).

MOTHER TONGUE

Issue 22, May 1994

REFERENCES

Bengtson, John D. 1990. "An End to Splendid Isolation: The

Macro-Caucasian Phylum." Mother Tongue 10 (April

1990).
Bengtson, John D. 1991a. "Notes on Sino-Caucasian," in

Shevoroshkin (ed.) 1991, pp. 67-129.
Bengtson, John D. 1991b. "Some Macro-Caucasian

Etymologies," in Shevoroshkin (ed.) 1991, pp. 130-141.
Bengtson, John D. 1991c. "Macro-Caucasian Phonology (Part

I)," in Shevoroshkin (ed.) 1991, pp. 142-149.
Bengtson, John D. 199 Id. "Postscript I. (Macro-Caucasian

Again)," in Shevoroshkin (ed.) 1991, pp. 150-156.
Bengtson, John D. 1991e. "Macro-Caucasian: A Historical

Linguistic Hypothesis," in Shevoroshkin (ed.) 1991, pp.

162-170.
Bengtson, John D. 1993. "The Macro-Caucasian Hypothesis."

Dhumbadji! 1/2:3-6.
Blazek, Vaclav and John D. Bengtson. 1995. "Lexica Dene-

Caucasica." Central Asiatic Journal (forthcoming).
Bouda, Karl. 1950. "Die Sprache der Buruscho." Eusko-

Jakintza 5:37-50 and 337-346.
Bouda, Karl. 1954. "Burushaski Etymologien." Orbis

3/1:228-230.
Cirikba, V. A. 1985. "Baskskij i severokavkazskie jazyki," in

Drevnjaja Anatolija, pp. 95-105. Moscow: Nauka.
Dolgopolsky, A[aron] B. 1964. "Gipoteza drevnejsego

rodstva jazykovyx semej Severnoj Evrazii s

verojatnostnoj tocki zrenija [A Probabilistic Hypothesis

Concerning the Oldest Relationship among the Language

Families of Northern Siberia]." Voprosy Jazykoznanija

2:53-63. (English translation in Shevroshkin and Markey

1986, pp. 27-50.)
Dumezil, Georges. 1933. Introduction a la grammaire

comparee des langues caucasiennes du nord. Paris.
Nikolaev, Sergei L. 1991. "Sino-Caucasian Languages in

America," in Shevoroshkin (ed.) 1991, pp. 42-66.
Sapir, Edward. 1913. "Wiyot and Yurok, Algonkin

Languages of California." American Anthropologist

15:617-646.
Sapir, Edward. 1915. "The Na-Dene Languages: A

Preliminary Report." American Anthropologist 17:534-

558.
Shafer, Robert. 1963. "Eurasial." Orbis 12:19-44.
Shafer, Robert. 1965. "The Eurasial Linguistic Superfamily."

Anthropos 60:445-468.
Shevoroshkin, Vitaly (ed.). 1991. Dene-Sino-Caucasian

Languages. Bochum: Brockmeyer.
Shevoroshkin, Vitaly and Thomas L. Markey (eds.). 1986.

Typology, Relationship, and Time. Ann Arbor, MI:

Karoma.
Starostin, Sergei A. 1982. "Praenisejskaja rekonstrukcija i

vnesnie svjazi enisejskix jazykov," in Ketskij sbornik, pp.

144-137. Leningrad.
Starostin, Sergei A. 1984. "Gipoteza o geneticeskix svjazj ax

sinotibetskix jazykov s enisejskimi i severnokavkazskimi

jazykami [On the Hypothesis of a Genetic Connection

between the Sino-Tibetan Languages and the Yeniseian

and North Caucasian Languages]," in Lingvisticeskaja

rekonstrukcija i drevnejsaja istorija Vostoka, part 4, pp.

19-38. Moscow. (English translation in Shevoroshkin

[ed.] 1991, pp. 12-41.)
Starostin, Sergei A. 1987. "Kommentarii k kavkazovedceskim

rabotam N. S. Trubeckogo," in Trubetzkoy 1987, pp.

437-473.
Starostin, Sergei A. 1989. "Nostratic and Sino-Caucasian," in

Vitaly Shevoroshkin (ed.), Explorations in Language

Macrofamilies, pp. 42-66. Bochum: Brockmeyer.
Tailleur, O. G. 1958. "Un ilot basco-caucasien en Siberie: Les

langues ienisseiennes." Orbis 7/2:415-427.
Toporov, V. N. 1971. "Burushaski and Yeniseian Languages:

Some Parallels". Travaux linguistiques de Prague 4: 107-

125.
Trubetzkoy [Troubetskoj / Tpy6en,KOH], Nikolai S. 1930.

"Nordkaukasische Wortgleichungen." Wiener Zeitschrift

fur die Kunde des Morgenlandes 37/2. (Russian

translation in Troubetskoy 1987, pp. 271-279.)
Trubetzkoy, Nikolai S. 1987. Izbrannye trudy po filologii.

Moscow.
Uhlenbeck, C. C. 1937. "The Indogermanic Mother Language

and Mother Tribes Complex." American Anthropologist

NS39.

The following article is reprinted with permission from
Diachronica VII/2.149-180 (1990).

INDO-EUROPEAN AND URALIC TREE
NAMES

LYLE CAMPBELL
Louisiana State University

1. INTRODUCTION

In this paper I present a number of similarities in tree
names shared by Indo-European (IE) and Finno-Ugric (FU) or
Uralic (U) languages. These shared terms are sufficient to
demonstrate some historical connection between the two
language families — due either to genetic relationship or
language contact, or perhaps both. 1 The purpose of this paper
is to present the evidence from tree names that the two language
families have a very old historical connection, whether genetic
or diffusional, and in doing so hopefully to stimulate further
research towards determining the exact nature of the
connection.

Both genetic relationship and areal (Sprachbund)
affinities have been proposed which include these two families.
Several proposals of distant linguistic relationship which link U
to IE have been made. The Indo-Uralic hypothesis continues
to be controversial, though attractive (see Collinder 1934,
1954, 1965; Joki 1973; Uesson 1970). Broader proposals of
remote genetic relationship are looked on favorably by some,
but skeptically by others. These include Nostratic (which,

MOTHER TONGUE

Issue 22, May 1994

whether in wider versions or narrower views, all include both
IE and U) (see Dolgopolsky 1984; Kaiser and Shevoroshkin
1988; Shevoroshkin 1989; etc.) and Eurasiatic (see Greenberg
1987 and forthcoming; Ruhlen 1987 and 1990), etc. Should
any of these proposals of remote genetic relationship prove
well founded, some of the shared tree names presented here
may turn out to be cognates.

On the other hand, research on possible distant genetic
relationships involving IE and U have tended not to take into
consideration the implications of areal linguistics. Various
proposals for linguistic areas (Sprachbunde) involving IE and
U, along with various other languages and families, have been
made (see, for example, Decsy 1988; Haarmann 1976;
Friedrich 1970; Jakobson 1938; Joki 1964 and 1973; Katz
1975; Sebeok 1950; Sinor 1969; Thomason and Kaufman
1988; Zeps 1962). Given these proposed linguistic areas and
documented early loans from IE languages found in FU
languages (cf., e.g., Joki 1973; Koivulehto 1976, 1980, and
1984; Korenchy 1988; Redei 1986-1988; Suhonen 1988; Tauli
1955), it is possible that some of the similarities presented here
may well turn out to be old loan words due to intimate contact
between PIE and PFU or PU languages.

It is hoped that the evidence assembled here from tree
names for an old connection between IE and FU (or U) may
help in further explorations of the competing (and perhaps
overlapping) explanations offered by distant genetic and areal
linguistic proposals for the similarities shared by IE and U
languages. In particular, given the evidence for diffusion
presented here, approaches to proposed distant genetic
relationships involving IE and U should be modified to include
the implications of areal linguistics. That is, while areal
linguistic considerations are given importance in remote
genetic research in other areas of the world, e.g., among
American Indian languages, in south and southeast Asia, and in
Australia, areal concerns have not played as important a role as
they should in proposals which attempt to connect IE (and FU
to a lesser extent) with other languages.

2. BACKGROUND

The subject of tree terms is of considerable
importance in the literature on both IE and U, since arguments
for the original homeland of each of the two families rely
heavily on the tree names (see below).

The classification of U languages is:

Uralic Family

I. Finno-Ugric Division

Population

1 . Early Balto-Finnic Group

a. Late Balto-Finnic

Finnish

5,000,000

Estonian

1,000,000

Karelian

138,000

Veps

16,000

Votic

a few dozen

Livonian

300

b. Lapp

38,000

2. Votic Group

Mordvin

1,200,000

Cheremis (or Mari)

622,000

3. Permic Group

Zyrian (or Komi)

478,000

Votyak (or Udmurt)

714,000

4. Ob-Ugric Group

Ostyak (or Khanty)

21,000

Vogul (or Mansi)

7,600

5. Hungarian

14,000,000

II. Samoyed Division

Population

1. Northern Branch

Nenets (or Yurak)

25,000

Enets (or Yenisei Samoyed) 400

Nganasan (or Tavgi)

1,000

2. Southern Branch

Selkup (or Ostyak Samoyed) 4,300

(After Korhonen 1984 and Hajdu 1975a, b) 4

3. TREE NAMES AND HOMELAND RESEARCH

The linguistic study of the Finno-Ugric (and Uralic)
homeland has an ample history, and there have been several
different opinions concerning its location and even its nature
(see E. Itkonen 1966:22-31). Tree names have played a very
significant role in this research. All the linguistic evidence for
the Finno-Ugric homeland points to an area located in the
Middle Volga or in the area between the Volga and the Urals
(Korhonen 1984:61; T. Itkonen 1984:350). Proto-Finno-Ugric
vocabulary offers limited clues for delimiting the homeland,
and tree names have been crucial.

The five principal trees involved are:

1. "spruce" [Picea obovata], PU **kdxsi, PFU *koosi
(Sammallahti 1988:538), with cognates in Balto-Finnic,
Lapp, Mordvin, Cheremis, Votyak, Zyrian, Ostyak, Vogul,
Nenets, Enets, Nganasan, Selkup, and Kamas.

2. "Siberian pine" [Pinus sibirica], PU **siksi, PFU *siksi
"cedar" (Sammallahti 1988:540); PU **sikse "Siberian
pine (Pinus sibirica)" (Hajdu 1979:37); with cognates in
Votyak, Ostyak, Vogul, Selkup, and Kamas (see below).

3. "Siberian fir" [Abies sibirica], PFP *nulka "silver fir"
(Sammallahti 1988:553); PU **nulka "Abies sibirica"
(Hajdu 1975a:37); with cognates in Votyak, Zyrian,
Ostyak, Vogul, Selkup, and Kamas.

4. "Siberian larch" [Larix sibirica], PFU *nar)e "Siberian

larch" (Hajdu 1975a:37); with cognates in Zyrian, Ostyak,
and Vogul.

5. "brittle willow" [Salix fragilis]/ "elm", PFU *siliw "elm"

MOTHER TONGUE

Issue 22, May 1994

(Sammallahti 1988:549); PFU *sala "elm, brittle willow
(Salix fragilis)" (Hajdu 1979:38; Collinder 1960:413),
with cognates in Finnish, Mordvin, Cheremis, and
Hungarian; outside of Finnish, the cognates mean "elm"
[Ulmus]. (Cf. Hajdu 1969:257 and 1975a:37-40).

The ancient distribution of these trees has been studied
from their pollen in layers of peat and silt from the bottom of
swamps. All except the "elm" have been part of the Siberian
flora from the earliest times. Nevertheless, the "spruce" spread
already during the Early Holocene over the Urals as far as the
White Sea. The "Siberian pine" and "Siberian fir" appeared in
Europe in the Middle Holocene (6,000 - 500 B.C.), but only in
the eastern part bordering the Urals, the upper reaches of the
Petsora; the "Siberian fir" also reached the Kama River. The
"Siberian larch" spread from the Urals to the west only around
500 B.C., but it was found in western Siberia already in the
Middle Holocene. The "elm", on the other hand, with its
western origin, began to spread in the Early Holocene from the
deciduous forests of Central Europe to the east and met the
Siberian conifer forest on the west side of the Urals in the
Middle Holocene. According to Hajdu (1964, 1969, 1975b),
the Finno-Ugric homeland could be located only in an area
where all these trees were found. The only place which fits
temporally and geographically is from the Middle Urals to the
north, including the lower and middle course of the Ob and the
headwaters of the Petsora in the area of the northern Urals.

Evidence for the original homeland has also been
sought in contacts with other languages. Proto-Finno-Ugric
has a significant layer of loans from Indo-European,
specifically Indo-Iranian. The contacts that led to these loans
could scarcely have happened anywhere else than in eastern
Europe (cf. Joki 1973). One such example is "honeybee".
From the beginning (Koppen 1886, 1890), the words for
"honeybee" (*mekse: Finnish mehi-ldinen, Mordvin m'eks,

Votyak mus, Zyrian mos, Hungarian meh) and "honey" (*mete:
Finnish mesi, Lapp mieta, Mordvin m'ed, Cheremis mix,
Votyak mu, Zyrian ma, Hungarian mez) have been emphasized
in studies of the FU homeland. These terms were borrowed
into PFU from IE, apparently from Proto-Indo-Iranian. The
area in which such contact could have taken place was the
region of the middle course of the Volga River, where
apiculture was practiced in one form or another from the
earliest times (Korhonen 1984:61; cf. Hajdu 1975b:33). 5

To sum up the Urheimat considerations, the linguistic
evidence available strongly suggests that the Proto-Finno-
Ugric (and Proto-Uralic, as well) original homeland was
located in the area somewhere between the Volga and the
Urals, and may have extended a bit over the east side of the
Urals. Nevertheless, Finno-Ugric-speaking people may have
come far to the west, as far as the eastern Baltic region, before
the break-up of the unified proto-language (Korhonen 1984; cf.
T. Itkonen 1984:350; Sammallahti 1984). It is hypothesized
that Proto-Uralic was spoken in the Middle Holocene when the
spruce, Siberian pine, and Siberian fir had already appeared in
Europe, or at least in the area of the Urals, i.e., after 6,000 B.C.
On the other hand, the Proto-Finno-Ugric period could not
begin before the elm's spread had reached the Siberian conifer

forest's western edges, ca. between 6,000 and 4,000 B.C.
(Hajdu 1964 and 1969). 6

It will not be overlooked that this postulated FU
homeland is quite near that assigned to PIE speakers in most
proposals. It is not surprising, then, that FU (also U) and IE
speakers would have known the same trees, although it is more
significant that their names for these trees are in several cases
quite similar. With this background in mind, I now turn to a
comparison of tree names in these two language families.

4. COMPARISON OF TREE NAMES 7

4.1. Cedar

Friedrich (1970:149) lists as possibly connected and
worthy of further study the IE forms:

Greek

Ke8poc;

Lithuanian

kadagys

Latvian

kadegs (SKES 170)

Old Prussian

kadegis "juniper" (Redei 1986

1988:165)

It seems, however, that these are borrowed from FU languages,
at least the Lithuanian and Latvian forms (SKES 170). They
can be compared to the FU forms:

PFU necrje (Collinder 1960:129 and 140), PFU
*kad3 "juniper" (Redei 1986-1988:133)

Finnish

kataja

Veps

kadag

Vote

kataga

Estonian

kadak(as), kadaj(as)

Livonian

kadaG

Zyrian

kac-

Proto-Lapp

*kesnes (kaskas, gasskas, gasnes)

(Lehtiranta 1989:42)

Vogul

kastaj-, kase-

Ostyak

kocrji, hasrja

Selkup Samoyed keca, kyca

Kamas Samoyed keedamgs

Koibal (Samoye

d) kaduma

(Collinder 1960:129; SKES 170)

The borrowing probably is from the Balto-Finnic
languages; it is uncertain whether these Balto-Finnic forms are
actually cognate with the others listed for which Collinder has
reconstructed the PFU form listed here. It is conceivable,
nevertheless, that these later cognates and the Greek form may
also have connection, although not as recent (and by no means
as obvious) as the Baltic borrowing from Balto-Finnic.

4.2. Willow

One IE term for "willow" is reconstructed as *sVlyk-
(Friedrich 1970:53), *sal(i)k- (Watkins 1985:56), with

MOTHER TONGUE

Issue 22, May 1994

cognates:

Greek

Latin
Old Irish
Mid. Welsh
Old English
OHG
Old Norse
Proto-Gmc.
Anatolian (?)

e^ncobv, eXikt] (in Arcadian, 6th

century B.C. F eXnccov)

salix

sail

helygen

welig

salah(a)

selja

*solk-

uelku- "grass"

The *salyk- cognates are in Western Indo-European;
to accommodate the forms with w, Friedrich posits a western
and central *s/wVlyk-. Also, the "willow" forms may be
etymologically related to "grass" (Friedrich 1970:55 and 155).

There are two FU cognates sets which compare well
to these IE forms; the first is:

PFU *siliw, PFP *selv, PUg *siliw "elm"

(Sammallahti 1988:549);
PFU *sala "elm, brittle willow (Salix fragilis)" (Hajdu

1979:38; Collinder 1960:413);
PFU Hala "elm" (Redei 1986-1988:458):

Finnish salava "willow" (dialectal halava)

Mordvin sel'er), sel'ej; sal'i "willow"

Cheremis so(-)l "elm"

Hungarian szil "elm", szilas "elm grove"

(SKES 954; Hajdu 1969:257 and
1975a:38; Collinder 1960:413)

This FU set is strikingly similar to the IE etymon, though the
FU word probably originally meant "elm", which shifted to
"willow" in Finnish (E. Itkonen 1969:303). The "elm", with its
western origin, began to spread in the Early Holocene from the
deciduous forests of Central Europe to the east and met the
Siberian conifer forest on the west side of the Urals in the
Middle Holocene (see above; Hajdu 1969 and 1975b).

A comparison of the IE "willow" set with the FU
forms should not occasion too much skepticism, since, in spite
of the "elm" meaning of the FU forms, "willow" is clearly
involved in the cognate set. That is, given the semantic shift
within FU involving the meanings of "elm" and "willow", a
similar semantic shift in some of the forms compared between
FU and IE is not implausible.

The other U etymon which can be compared with the
IE form is PFU *silka, PUg *sila "pole" (Sammallahti

1988:549); PFU *salks (Collinder 1960:413); PFU *salka
"pole, stick, staff (Redei 1986-1988:460):

Finnish salko "thin, long pole, staff, stick,

stake"
Mordvin salgo "stick, pin"

Vogul saila, saila "slat, lath, splint"

Ostyak say5t, saysl, sa*x51 "splinter, lath,

picket"

(SKES 955-956; Collinder 1960:

413)

While semantically a comparison of "willow" with
"thin pole" is not as compelling as semantically equivalent
forms would be, these are, nevertheless, sufficiently similar in
meaning to invite curiosity about possible connections. 9

4.3. Poplar, aspen

PIE *osp- (Friedrich 1970:49) is attested in three
subgroups (cf. Watkins [1985:3] *apsa "aspen"):

Old English aespe "white poplar, aspen"

North Sorbian wosa "white poplar"

Lithuanian apuse, epuse "black poplar"

Latvian apse "quaking asp"

Russian osina "quaking asp"

Proto-Slavic

"opsa aspen

To these, the following additional but problematic IE forms can
be compared (Friedrich 1960:50-51):

Armenian

op'i "white poplar" (fit not

explicated)

Greek

aoniq "shield" (improbable cognate)

Sanskrit

sphya- "front oar on a boat, boat

pole, shoulder blade, sacred

instrument"

Wakhi (Iranian)

fiak "shoulder"

Persian

fah "oar, paddle"

Proto-Indo-Iran.

*sphiya-"oar, shovel, shoulder

blade"

There are several interesting aspects of this etymon,
even with attention restricted to the first, less controversial set
of forms. Indo-Europeanists disagree about the sequence of
medial consonants, and some think the word looks non-Indo-
European (Friedrich 1960:49). Friedrich points to similar
forms for "Populus" in several languages of southern Siberia:

Tobol

ausak

Altai

apsak

Lebed

apsak

Kumandu

aspak

Chuvash

eves

Sagai

OS, OS

From this evidence, Friedrich (1970:50) concludes:

during the third and second millen[n]ia both Indo-
European and non-Indo-European languages shared
forms with both orderings of medial consonants that
referred to poplars and aspens. Osp- was an areal term
shared by members of a Sprachgruppe of interacting
speech communities in eastern Europe and southern
Siberia.

MOTHER TONGUE

Issue 22, May 1994

Friedrich (personal communication, 1975) intends the
Sprachgruppe to include FU and IE; this would be clear in any
case based on the similarities exhibited in the FU names for this
tree. We can compare: PFV *sapa "aspen, poplar" (Redei
1986-1988:783):

Finnish haapa "aspen, poplar"

Proto-Lapp *supe "aspen, poplar" (Lehtiranta

1989:126)
Cheremis sa-pki, sa-pi, sopke- "aspen, poplar"

(SKES 46)

4.4. Another U "aspen/poplar" term

Uralic has another, more extensively attested etymon
for "poplar, aspen", which also bears considerable similarity to
IE and other forms already cited above. The reasons for
associating the IE and these U forms are quite obvious.
Nevertheless, the U data are a bit cumbersome. It seems that
originally there may have been perhaps two distinct but
phonetically similar etyma, one for "aspen/poplar" and another
for "willow", which have in some ways merged, either in the
actual languages or at least in the minds of those presenting the
forms. It becomes very difficult to separate them, if indeed
they were originally distinct. I present both here:

PU **poje "aspen, poplar (Populus)" (Hajdu
1979:37); PU **poja "aspen (Populus tremula)"
(Decsy 1990:106); PFP *pqji "aspen" (Sammallahti
1988:553); PU **poj3 "aspen" (Collinder 1960:408).
PU **pdjiw 9 PFP *pajiv, PUg *poji- "willow"
(Sammallahti 1988:548); PU **paje "willow (Salix)"
(Hajdu 1979:37); PU **paja "a kind of Salix" (Decsy
1990:105); PU **paj3 "willow species" (Redei 1986-
1988:349), PU **poj3 "aspen" (Redei 1986-
1988:391):

Finnish

Mordvin

Votyak

Zyrian

Ostyak

Yurak Samoyed

Ostyak Samoyed
Votyak

Proto-Samoyed
Yurak Samoyed
Selkup Samoyed
Kamas Samoyed

4.5. Conifer species

paju "willow"

poj, poju "aspen, poplar"

bad' "willow"

bad', baid "willow"

poi, pai "aspen, poplar"

p'e|3 "basket with willow inside"

p'e, py "willow basket"

pi(-pu) "aspen"

*pi "aspen" (Janhunen 1977:123)

pi "aspen"

■■ '« "10

pn-ni aspen

(SKES 465-466)

Friedrich (1970:34) reconstructs *pytw, based on these forms:

Greek tutix; "pine, spruce"

Latin plnus "pine, fir" (< *plt-snus or

*p!t-snos)

Sanskrit pitu-daru- "a kind of pine, fir, or

resinous tree"

(Some have suggested a Dravidian
origin for this, compare Telegu
pisunu, originally "gum, resin"
[Friedrich 1970:34].)

Albanian pishe "pine, fir" (related to

peshk(ve) "resin, tar") {pishe
perhaps from *plt-s-ia)

This reconstruction is supported by three
Mediterranean stocks and doubtful Indie. Benveniste (1955)
did not think it was part of the common IE lexicon. Friedrich
(1970:34) points to "alleged loans" into Balto-Finnic (e.g.,
Finnish petaja) "as possible evidence of its antiquity".
However, this form has a fairly respectable antiquity in FU,
going back at least to Proto-Finno-Permic times. Some forms
supporting this are:

PFP *peda "pine" (Sammallahti 1988:553); PFP
*peca\ *penca "Pinus silvestris" (Redei 1986-
1988:727)

Finnish

petaja "pine"

Proto-Lapp

*pece "pine" (Lehtiranta 1989:100)

Mordvin

p'itse "pine"

Cheremis

pandza, pundzo, penze "pine"

Votyak

puzim "pine"

Zyrian

pozem "pine"

(SKES 534)

Several terms for conifers will be considered. The
first is a "pine" term. As a "weak and chronologically late PIE
tree name denoting some sort of conifer, probably the pine",

Hajdu (1964:54 and 1979:37) finds a possible, though
uncertain Hungarian cognate fenyo "pine"; he reconstructs PFU
*pede "Pinus silvestris" ("pine"). Redei (1986-1988:416)
reconstructs PFU *pon3 "fir, spruce", based on such cognates
as:

Cheremis pin "zum Pflanzen gelassene

Fichte"
Votyak pumel' "fir"

Zyrian pomrl' "young fir, juniper"

Hungarian fenyo "fir, spruce, pine"

The fact that the form is only weakly attested in IE
lends support to those inclined to suspect a borrowing from FU.
However, the lack of clear and uncontroversial cognates
beyond Finno-Permic within FU would tend to sober too much
enthusiasm for a facile assumption of FU-to-IE borrowing. On
the other hand, forms similar to these from Turkic and other
so-called "Altaic" languages suggest wider areal diffusion,

F-

MOTHER TONGUE

Issue 22, May 1994

compare:

Yakut (Turkic)
Soyot (Tuva)

bas "Scotts pine"

pqjs "Pinus cembra" (Rasanen

1955:26)

The connections with "resin, pitch" (see below) may help to
sort out the history of these terms.

4.6. Second "conifer" connection

The evidence for historical contact between the two
families is strengthened by the close agreement among other
conifer terms. A second of Friedrich's IE conifer terms,
bearing a certain phonetic similarity to the first, is supported by
three stocks — *pyk- (Friedrich 1970:35), with the following
cognates:

Greek

7iiooa "resin, tar"

Latin

pix, picis "pitch, tar, pine"; picea

"spruce"

Umbrian

peiu "spruces" (< *pik-ie-)

Old Ch. Slavic

pi>cbh> "pitch, tar"

Serbo-Croatian

pakao "pitch, tar"

Polish

pieklo "pitch, tar"

These (as well as the Albanian forms in 4.5 above)
compare well with the FU forms:

PFU *piska "resin, pitch, gum" (Collinder 1960:413);
PFU *piska "resin" (Redei 1986-1988:385-386):

Finnish pihka "resin, pitch, gum"

Veps pihk "pitch, fir, dense forest"

Vote pihku "pine, pitch"

Ostyak p'lysA, piyli'ta ' "to plug with

pitch, to tar a boat"
Votyak pi'ylita "to plug with pitch, to tar a

boat" (SKES 541-542)

The similarity in these last two cases between the PIE
and FU forms would seem to be striking enough to make
chance an unlikely explanation.

4.7. Third conifer term

The reconstruction of a third IE conifer term is
supported by four subgroups. It is *pwK- or *pewk- "pine,

spruce", with cognates:

Greek
Lithuanian
Middle Irish

OHG

nevKt] "spruce, pine, spruce forest"
pusis "spruce, pine"
ochtach "kingpost of a house"
(< *ptdctata);

octgag "pine"
octhgacha "fir"
fiuhta, fiehta "spruce"

This set, with somewhat limited distribution within IE,
compares well to U terms for "tree", where in the U area
"spruce and "pine" are the most typical of trees:

PU **puxi/**poxi/**pdxi (final **/ or **f), PFU
*puxi, PS *pti, PU *pugi "tree" (Sammallahti
1988:539); PU **pu (Collinder 1960:408); PU
**punga "tree, wood" (Decsy 1990:106); PU **puwe
"tree, wood" (Redei 1986-1988:400):

Finnish

puu "tree, wood"

Cheremis

pu "tree"

Votyak

pu, pui "tree"

Zyrian

pu "tree"

Vogul

-pa "tree"

Hungarian

fa "tree"

Proto-Samoyed

*pa "tree, forest, wood" (Janhunen

1977:117)

Yurak

p'a "tree"

Yenisei

pe "tree"

Tavgi

fa "tree"

Ostyak Samoyed

i po, pu "tree"

(SKES 664)

There is also another U etymon which can be compared with
the IE forms, though I find it only in Collinder:

PU **piika "cone of conifer" (Collinder 1960:408)

Cheremis pugQl'mo "cone of conifer"

Ostyak poki "seed of a coniferous tree"

Selkup Samoyed puuga, puuka, puiiga "seed of a fir

tree"

(Collinder 1955:53)

4.8. Fourth "conifer" term

Hajdu (1969:257) cites another possible conifer
connection between U and IE. He gives IE *sukse- "pine
(Pinus cembra)", which he compares with the PU **sukse of
the same meaning. I am uncertain concerning this IE form, but
find the following IE and U cognate sets quite comparable:

PIE *s(u)ek~o-s "resin, sap", Watkins (1985:68)
*s(w)okwo- "resin, juice":

Greek

onoc, "resin, sap"

Latvian

svakas "resin, sap"

Lithuanian

sakaT "resin, sap"

Russian

osoka "Bluteiter"

Albanian

gjak "blood"

(Pokorny 1959:1044)

And:

PU **siksl, PFU *siksi, PS *tite 9 PFP *seksi, PUg

MOTHER TONGUE

Issue 22, May 1994

*6ik0i "cedar" (Sammallahti 1988:540); PU **sihe
"Siberian pine (Pinus sibirica) (Hajdu 1979:37); PU
**soksa/**seksd "Pinus cembra" (Decsy 1990:107);
PU **soks3 (**saks3, **seks3) "Pinus cembra"
(Redei 1986-1988:445):

Votyak

susi-(pu) "juniper"

Zyrian

sus-(pu) "cembra pine"

Vogul

tet, tat "pine" (t- < *s-)

Ostyak

texat

Yurak Samoyed tydo ?

Selkup Samoyed tyty

(Hajdu 1969:257) 1Z

.9. Fifth conifer

Another conifer listed under Friedrich's (1970:151)
"Miscellanica Arborea" as "pine, cedar" is illustrated with the
IE forms:

Armenian mair "pine, fir, cedar"

Latvian mltra "boxtree"

Old Ch. Slavic smrect "juniper"
smreca "cedar"

These can be compared with such FU forms as:

PFU *mor3 "a species of tree" (Redei 1986-1988:281)

Proto-Lapp *more "tree" (Lehtiranta 1989:78)

Hungarian mor "spruce"

As for other languages with possible areal connections, Tungus
dialects have mar "spruce" (Sinor 1 969:280). 1

4.10. Apple

Terms for "apple" seem to represent a true
Wanderwort in northern Eurasia. I will consider first the IE
and FU terms, and then those from other languages that seem
to be involved. Friedrich (1970:57) suggests two forms for IE
"apple", *abVl- and *maHlo-, saying that these rest "to an
unusual degree on external, nonlinguistic criteria". The first
has the supporting forms:

Pre-Celtic

*abalos "apple"

Old Irish ubull "apple"

Middle Welsh afallen "apple tree"
Old English aeppel "apple"

Pre-Baltic

*abul "apple"

Latvian abuolis "apple"

Old Ch. Slavic (j)abhJo> "apple"

Russian

jabloko "apple"

*abal-n- "apple
tree"

apuldor "apple

tree"

*abeles "apple

tree"

abels "apple tree"

(j)ablanb "apple

tree"

jablom> "apple

tree

On the basis of these forms, Friedrich (1970:58-59) posits a
North Indo-European *abVl-. This reconstruction is peculiar,
however, in that it has the otherwise very rare IE b and a
divergent /-stem.

Joki's (1963 and 1964) survey of "apple" terms in
various language families of the region (together with different
opinions held concerning these) is instructive here. It might be
noted that Sanskrit phdla-m "fruit" has been thought by some
to be cognate with the apple terms, while others have thought
the North Indo-European terms may be borrowed. Finally, the
Iranian forms have been much discussed (cf. Friedrich
1970:63; Joki 1963), as seen in:

Wakhi

miir, mirr "apple"

Zaza

meroe "apple"

Pahlavi

mur

Modern Persian

mul "poor wild pear"

Shughni

mun, mun "apple"

Sar(i)koli

man "apple"

Munjani

amirjga "apple"

Proto-Iranian

*amana/*amara "apple'

(Joki 1963:135)

Morgenstierne (as cited by Joki 1963:135) in his Afghan
etymological dictionary gave: mana, mana "apple", manu
"chokeberry (?)", and mara una (pi.) "sour apple", and derived
all from *marna-. In addition to those already cited above, he
compares:

Sarkoli

man "apple-

Ishkashmi

mind "apple"

Jidghan

amunoh "apple"

Parachi

amar "apple"

Ormuri

millz, mile "apple'

The following FU terms can be compared with these
from IE, as well as with the forms in other languages to be
presented below:

PFV *omena ~ *om3r3 "apple" (Redei 1986-
1988:718)

Finnish omena "apple" ("pear" also in

dialects)
Estonian omin, ubin, 6un(a) "apple"

Livonian umar "apple"

Mordvin umal, mal "apple, fruit, berry"

mafl'u "apple tree"

Votyak umo "apple" (probably from

Chuvash) (Redei 1986-1988:718)
(SKES 429-430)

Some have suggested that FU may have borrowed
from Old Iranian or the "Scythians" (Joki 1973), while the
borrowing may also have been in the other direction. However,

MOTHER TONGUE

Issue 22, May 1994

the picture becomes more interesting when the forms for the
following additional languages from northern Eurasia are taken
into account:

Chuvash olma, omla, uma "apple"

Mongolian alima, aliman, alim "apple"

Rasanen (as reported by Joki 1963:135) thought these were
from *alyma with a prothetic *a- and that they were perhaps
borrowed from Sino-Korean *//m. It is usually assumed that
Votyak *umo "apple" is borrowed from Chuvash (SKES 430).

Basque terms have also been compared: mandaka
"sour apple species" (< *mantar)ka), mandaitu "a kind of
apple", amun "a kind of apple" (< *a-mont) (Joki 1963:135).

Joki investigated all these "apple" terms in detail,
concluding that the Turkic and Mongolian "apple" terms are
also cultural loans spread widely from one language to another,
whose probable origin was the fruit areas of or near Tajikistan
(Joki 1963). Since there has been such widespread borrowing,
and since the reconstructions are not undisputed in any of the
language families involved, we may never know in each case
which languages borrowed from which others, or whether any
may share the form as a cognate from a common parent
language.

4.11. The second apple term

Friedrich's (1970:60-62) second IE "apple" form,
which is disputed, is attested in the following:

Doric Greek |i&A,ov "fruit, apple" (poetic

"cheek")
Homeric Greek uf]A,ov "fruit, apple"
Latin malum "apple, fruit"

Albanian molle "apple (apple-cheeks)"

(possibly a loan from Latin)
Hittite mahla- "grapevine, apple(-tree)" ?

Tocharian A malafi "cheeks" (metaphorical

shift ?)
Cf. Hittite mu-ri-es "grape"

The probable relation of "apples" to "grapes" and to "fruit" in
IE raises other interesting possible connections. In this respect,
we find several U forms which offer suggestive similarities. In
what follows, I list these with some discussion where helpful.

(1) "Berry" 1.

PU **murd, PFU *murd, PS *merd "berry"
(Sammallahti 1988:538); PU **mora "raspberry,
cloudberry" (Collinder 1960:407); PU *mura
"Rhubus chamaeomorus" (Decsy 1990:103); PU
**mura "Rhubus chamaeomorus" (Redei 1986-
1988:287).

Finnish muurain, muurama "cloudberry,

arctic raspberry"
Zyrian mir- "cloudberry, arctic raspberry"

Vogul morax "cloudberry, arctic

raspberry"
Ostyak mur§x "cloudberry, arctic

raspberry"
Yurak Samoyed mararja "cloudberry, arctic

raspberry-
Yenisei Samoyed moragga "cloudberry, arctic
raspberry"
(SKES 355)

The Uralic cognate set bears a strong similarity to
Latin morum, Greek ^topov "mulberry", which Friedrich
(1970:150) links as possible cognates worthy of further study.
Both these and the "apple" terms are similar to the next U set.

(2) "Berry" 2.

PFU *marfa "berry" (Collinder 1960:412; cf. Redei
1986-1988:264):

Finnish marja "berry"

Mordvin mal "apple" (Moksha), "berry"

(Erza)
Cheremis mor, -mcrrs "berry"

Vogul moari, morip "berry cluster"

Ostyak murap', muirap "berry bunch"

(SKES 334)

The nature of the relationship between these last two cognate
sets is unclear — they appear to be distinct, yet to have a certain
overlap.

(4) "Berry" 3.

PFU *medi "berry" (Sammallahti 1988:544); PFU
*mel3 "cherry" (Collinder 1960:412); PU **mat3 "a
berry species" (Redei 1986-1988:265), PFU *mol'3 "a
berry of a certain shrub" (Redei 1986-1988:279):

Zyrian mol' "berry"

Ostyak *mel "berry"

Votyak mul'i "berry"

Hungarian meggy "cherry"

4.12. Beech

For "beech", Friedrich (1970:106) reconstructs PIE
*bhago-; Watkins (1985:5) gives *bhago-. However, Indo-

Europeanists have proposed reconstructions with five different
vocalisms (Friedrich 1970:106). Some of the IE cognates are:

Mod. Icelandic baukr "mug, box"

beyki "beech forest"
(probably invalid cognates)
OHG and MHG buohha "beech woods"
OHG, Old Norse bok "beech" (o < PIE *a)

MOTHER TONGUE

Issue 22, May 1994

Mod. English

beech

MHG

butchen "to wash or soak in hot lye

or buck"

Old English

buk "buckets"

(Old English and MHG are not real

cognates)

Thraco-Illyrian

mugo, muso "beech"

Lydian

musos "beech"

moesia "beech forest"

Old Ch. Slavic

frbzij "elder"

Russian

buzina "elder"

Ukrainian

byze "elder"

Albanian

bunge "Quercus esculentis"

Serbo-Croatian

baz "elder"

Czech

bez "elder"

bzina "elderberries"

Latin

fagus "beech"

Celtic

bago "beech" (the Celtic form is

complicated because in some of the

languages the word has been

borrowed from Latin, leaving the

status in the others uncertain)

Doric Greek

cpayoq "Quercus esculentis"

Attic Greek

cpnyoc; "Quercus esculentis" (with

edible acorns)

Beeches do not grow in the Baltic region nor in most of the FU
area. This accounts for why Baltic languages have no reflexes
of *bhago-, and perhaps for why the Slavic languages have
shifted the meaning to "elder".

Though the beech is not found in most of the FU area
(Friedrich 1970:113-114), its importance for "nuts" and the
semantic shifts to "elder(berries)" and "oaks (with edible
acorns)" suggest a possible connection with FU terms for
"nuts":

PFP *paski "nut" (Sammallahti 1988:553); PFP
*paste "nut, hazelnut" (Redei 1986-1988:726)

Finnish pahkina "nut" (dialectal "acorn")

Mordvin p'est'e, p'estse "nut" (st < *sk)

Cheremis piiks "nut"

Votyak pas-pu "nut bush" (-pu = "tree")

pas-mul'i "nut" {-mul'i = "berry")

(SKES 679)

Although the FU cognates extend this term only to Proto-
Finno-Permic times, it may be older, with no surviving
cognates in the other languages. 15

Another cognate set, which is phonetically similar but
semantically not so equivalent, can also be compared with the
IE "beech" forms. It is the "rowan" (mountain ash), which also
produces berries. Sammallahti (1988:553) reconstructs it only
to Proto-Finno-Permic, but SKES (542) and Collinder
(1960:413) finds it throughout FU:

PFU *picla "mountain ash, rowan" (Sorbus

aucuparia) (Redei 1986-1988:376); PFP *pisla
"rowan (mountain ash)" (Sammallahti 1988:553);
PFU *picla "rowan" (Collinder 1960:413)

Finnish

pihlaja "rowan"

Mordvin

p'izol "rowan, rowan berry"

Cheremis

psza-lma, pi-zls "rowan, rowan

berry"

Zyrian

pelys "rowan berry"

Votyak

pales "rowan berry", pales-pu

rowan

Ostyak

paDa'r, paza'r "rowan berry"

Vogul

pit'sa-r, paser "rowan berry"

Hungarian (?)

fagyal "honeysuckle"

(Sammallahti 1988:553; SKES 542)

Rasanen (1955:27; cf. also SKES 542) also pointed
out similarities in some Turkic languages, with the possibility
of borrowing from FU into Turkic:

Chuvash
Kazan Tatar
Turkish
Tobol

piles "rowan"

milas "rowan"

meles "rowan"

micar "rowan" (this Tobol form was

believed to be a loan from Vogul,

cf. Rasanen 1955:27)

4.13. Yew

Friedrich (1970:121) reconstructs IE *eywo- "yew",
although it means "yew" consistently only in Germanic and
Celtic; it is based on the cognates:

Greek

6a, oir| "service tree, bird-cherry"

Latin

uva ""grape, bunch of grapes, laurel

berry"

Armenian (?)

aigi, oigi "grapevine"

Lithuanian

ieva "black alder"

Latvian

iva "bird-cherry"

Old Prussian

iuwis "yew"

Czech

jiva "yew"

Serbo-Croatian

iva "willow" (the majority forms

mean "willow")

OHG

Twa (German Elbe) "yew"

Old Norse

yr "bow"

Breton

ivin "yew"

Gaulish

ivo- "yew"

Middle Welsh

yw-en "yew"

Old Irish

eo, i "yew" (< *iwo~)

ibhar "bow, yew"

The U forms for "yew" are strikingly similar:

PU **/oxf "tree", PFU *juxi 9 PS *je, PUg *jugi "tree"
(Sammallahti 1988:537); PU **juw3 "yew"
(Collinder 1960:406); PU **juwe "Pinus silvestris"

MOTHER TONGUE

Issue 22, May 1994

(Hajdu 1979:37); PU *juva "Pinus silvestris" (Decsy
1990:99); PU **juw3 "Pinus silvestris" (Redei 1986-
1988:107)

Finnish

juko(-puu) "yew" (-puu = "tree")

Estonian

juka(-puu), juha(-puu) "yew"

Cheremis

jakte "high-trunked tree, pine,

spruce, birch"

Ostyak

jux "tree, wood"

Vogul

jiw "tree, wood"

Proto-Samoyed

*je, *jew "pine" (Janhunen

1977:42)

(Hajdu 1964:54; Collinder 1960:

406; SKES 122)

4.14. Oak

There are several terms for oaks in IE. Friedrich
(1970:140) points out that cognates for *dorw- "tree, oak" are
attested in all the stocks (*derw-, *drw- also). Some of the
cognates are:

Hittite
Sanskrit

Avestan

Modern Persian
Greek

Latin

Lithuanian

Latvian

Old Ch. Slavic

Russian

Armenian

Old English
Albanian

Old Irish

Cornish
Middle Welsh

taru "wood"

daru "piece of wood, wood, wooden

implement"

druma- "tree" (in later texts)

drunam "bow"

draos "of the wood"

drvo 'hale"

daruna "rainbow"

56pi) "beam, wood, spear"

5po(f )dg "wood nymph"

5pCq, 5puoc; "oak (?), tree"

larix (?) "larch"

durus "hard"

darva "tar, log, piece of pine wood"

drutas "strong, stout"

darva "tar"

drevo "tree"

derevo "tree"

targal "spoon"

tram "firm, fixed"

treo(w) "tree"

dru, drum" "beam, piece of wood,

post, tree"

dry, dryni "wooden peg"

drochta "(wooden) trough"

derucc "acorn"

doar "oak"

dar "oaks"

derw-en "oak"

As Hajdu (1964:57) has pointed out, there are no "Quercus"
forms which extend to FU times. Only if the oaks had a wider
distribution in FU times would we expect terms for them in the
proto-language. Nevertheless, there are some terms in FU
languages which bear remarkable similarities to this IE etymon.
These are presented in the following examples:

(1) "Acorn": Hakulinen (1968:254) finds this term as
far back as Proto-Finno-Permic or Proto-Finno-Volgic times,
though this is unconfirmed:

Finnish

terho "acorn"

Vote

turu, toro "acorn'

Estonian

toru, toro "acorn'

Livonian

te'rm§z "acorn"

(SKES 1273)

Given the limited distribution among FU languages and the
striking phonetic (and semantic) similarity to IE forms, one
would suspect borrowing from IE.

(2) "Fir, pine": Another possibility is:

PFP *tirka "fir" (Sammallahti 1988:550)

Cheremis terke "pine" (Hajdu 1964:56)

Vogul tareg "fir"

Ostyak teeger "fir"

(3) "Oak": Hajdu (1964:56) has given a possible, but
doubtful FU cognate set for "oak":

Hungarian tolgy "oak" (< *til-)

Ostyak tal "forest"

Zyrian tel' "wood of a Nagelbaum"

4.15. Another "oak"

Among his miscellaneous tree terms which warrant
further study, Friedrich (1970:150) lists the following
additional "oak" term:

Proto-Germanic *danwo ("oak" in one 1 lth-century
gloss, but generally "pine" in OHG)

Old Ch. Slavic dgbt "oak"

Sanskrit dhanu-h "bow"

dhanvana-h "fruit tree"

One etymon found among Finno-Ugric languages for terms for
"oak" bears a similarity to these IE forms:

PFV, PFP (?) *toma "oak" (Redei 1986-1988:798)

Finnish tammi "oak"

Mordvin (Erza) tumo "oak"
Mordvin (Moksha) tuma "oak"
Cheremis turn, tu-ma "oak" (SKES 1218)

Votyak ti-pi, ta-pa (-pi, -pd = "tree")

Zyrian tu-pu (-pu = "tree") (Redei 1986-

1988:798)

A different, more extensively attested U etymon can also be
compared to this set of IE forms:

MOTHER TONGUE

Issue 22, May 1994

PU **d'ixmi, PFU *d'iimi, PS *je0m "bird-cherry"
(Sammallahti 1988:536); PU *dhjoma "Primus
padus" (Decsy 1990:99); PFU *8'ome "bird-cherry"

(SKES 1408); PU **ffeme 9 **&ome "bird-cherry"
(Redei 1986-1988:65)

Finnish tuomi "bird-cherry"

Proto-Lapp *5ome (Lehtiranta 1989:32)

Mordvin (Moksha) lajrhe
Mordvin (Erza) Tom

lomp§, lombo (< lorn +pu "tree")

l'§m

Tern

Tern, l'am

Cheremis

Votyak

Zyrian

Vogul

Ostyak Samoyed feu, f sem "bird-cherry"

(SKES 1408)

4.16. Linden

For IE trees connected with "linden", Friedrich
(1970:88) considered *lenTa, *leipa 9 llwyfen. As pointed out
by Friedrich (1970:89), the linden was both technologically
and ritually important to both IE and FU speakers.
Nevertheless, the IE status of the tree names is not clear. Some
compared forms include:

Greek oXi^okoq- "a species of oak" (?)

Celtic Lemo-, Limo- (in place names)

Middle Welsh llwyfen "elm, linden"

Balto-Slavic *leipa "linden"

Lithuanian liepa "linden"

Latvian liepa "linden"

Common Slavic *lipa "linden"

OHG linta "linden"

Icelandic lind "linden"

English linden

Lithuanian lenta "board"

Albanian (Tosk) lende "wood, material"

These can be compared with two different U etyma:

PFV *lems3, *leme-s3 "young linden" (Redei 1986-
1988:688)

Finnish lehmus "linden"

Mordvin l'evs "linden bast"

Cheremis nemests "young linden"

(Redei 1986-1988:688)

And:

PFV *leppa "alder" (Redei 1986-1988:689)

Finnish leppa "alder, birch"

Lapp lei'pe, liahp "alder"

Mordvin

4.17. Hornbeam ?

l'epe, l'epa "alder"
(Redei 1986-1988:689)

Friedrich (1970:99-106) considers IE cognates which
support his reconstruction *grobh- "hornbeam" (genus
Carpinus), of which he says, the wood is heavy, hard, and
elastic, but is "ideal for tools, weapons, and armor" (Friedrich
1970:99). Terms for "hornbeam" were used for "beech" in
some IE dialects (Friedrich 1970: 101). Some of the forms cited
by Friedrich are:

Old Prussian wosigrabis "Carpinus betulus"
Lithuanian skruoblas "hornbeam"

Latvian Gruoblas (a place name)

Ukrainian hrab "common hornbeam"

(Carpinus betulus)
Greek YP&|3- "a type of oak"

These can be compared to FU forms with which they bear a
certain (though not overly compelling) similarity:

PFU *kapp3 "a species of tree which can be worked:
poplar, aspen, fir"

Votyak
Ostyak
Vogul

4.18. Some loans

kipi "log, block"
xap "aspen, boat"
gap- "fir, poplar"
(Redei 1986-1988:12)

I finish this paper by briefly mentioning some cases of
known loans in FU from IE involving terms for trees.

(1) The Balto-Finnic word for "tar" is a loan
involving the IE *dorw- "oak, tree" etymon. It is believed to
be borrowed either from Baltic (Lithuanian darvd "tar", dervd
"pitch, tar, resinous pine tree"; Latvian darva "tar") or from
Germanic *terwon "tar" (SKES 1276):

Finnish

terva "tar

Estonian

torv "tar"

Livonian

tera "tar"

(2) Another fairly clear loan involves other "oak"
forms. In this case, it also shows how semantic shifts can be
involved in the transfer of tree names. Borrowed from Proto-

Scandinavian *aik- "oak'
1970:132]) are:

Finnish

(cf. IE *ayg- "oak" [Friedrich

aihki (dialectal aikki) "tall pine, tall

spruce"
Proto-Lapp *ajkke "ancient tree"

Lulea Lapp hai'hka "large pine"

Norwegian Lapp hai'ka, ai'ka "large pine, a tree

which grows in the south" ("oak"

MOTHER TONGUE

Issue 22, May 1994

only in fairy tales)

5. CONCLUSION

(3) Another set of loans involves the religious
connotations associated with IE *perk w - "oak" (Friedrich
1970:133). As Friedrich points out, this oak term had various
religious associations; cognates in Indie refer to "rain cloud,
god of storm, paradise tree, sacred fig tree, mountain". Slavic
tribes had perum* "the thunder god", while Celtic and Baltic
had similar deities. Germanic forms mean "forest, mountain,
tree, man, thunder god, and mother of Thor". Finnish has the
borrowing perkele "devil, satan" (obscene, used for cursing)
from Baltic; cf. Lithuanian perkunas, Latvian perkuonis
"thunder, thunder god". Other Balto-Finnic languages have
related forms, e.g., Vote perku "hell", Estonian pergel, porgel
"devil", porgu "hell". Erza Mordvin Pufgine "thunder" is also
borrowed from Baltic. Finnish piru "devil, satan" (also an
obscene curse word) may be from Old Russian peruwb
"thunder god" (cf. SKES 523-552, 576-577; Friedrich
1970:113-140).

(4) Friedrich (1970:55) reconstructs a second IE
"willow" term, *wyt-, based on the cognates:

In summary, the two language families, FU (or U) and
IE, share a rather large number of similarities among their
names for trees. While conceivably some of these compared
forms are but fortuitously similar, the weight of the aggregate
of comparisons is sufficient to support the conclusion that these
two language families have a very old historical connection,
one which reflects either a genetic affiliation or Sprachbund
affinities, or perhaps both. The connection may involve
diffusion, and indeed, certain of the forms presented here
almost certainly involve borrowing. Given this set of
circumstances, the possible areal linguistic relationship to
explain these and other observed similarities shared by
languages of these two families deserves serious consideration.
On the other hand, at this stage we cannot rule out the
possibility that the similarities among some of the tree names
explored here may perhaps reflect an old genetic relationship
— a common ancestor. Further study ought to keep both
hypotheses open, and it is hoped that the comparisons
presented here will indeed stimulate further research aimed at
determining the exact nature of the historical connection shared
by IE and FU (or U).

Russian

vif "woven object"

vetla "willow"

Lithuanian

vytis "willow switch"

Latin

vitis "willow, grape, tendril, shoot'

Greek

8iT8(F)a "willow, willow shield"

oioog "willow species"

Irish

feith "fiber, cord"

Old English

wldig "willow"

Old Norse

wit>ir "willow"

Sanskrit

vetasa-h "willow, rod, switch"

Avestan

vaetay "willow, switch"

Armenian

(?) gi "juniper"

In all nine IE stocks, the terms refer not only to "willow", but
also to "willow withes" or "shoots" and to artifacts made from
them (Friedrich 1970:54). These can be compared to the Balto-
Finnic terms:

Finnish

vitsa "switch, rod"

Vote

vittsa

Estonian

vits "twig, switch"

Livonian

vitsa

(SKES 1799)

These and the IE terms appear to be related in some
way. SKES (1799) tentatively attributes the Finnic forms to
borrowing from Russian; Hakulinen (1968:254), on the other
hand, gives these a more native origin, finding cognates for
them that go back to Finno-Volgic or Finno-Permic times.
SKES (1977) does not, however, believe the Latvian forms
vica, vice "switch" to be native, attributing them to borrowing
from either Estonian or Russian. Friedrich (personal
communication, 1973) holds it more likely that the Finnic
forms are loans from Slavic.

NOTES

1) While some may be accidental similarities, the
nature and quantity of the aggregate argue strongly that chance
is not the explanation for most of them.

The abbreviations used in this paper are:

Finno-Ugric

Indo-European

MHG

Middle High German

OHG

Old High German

PFP

Proto-Finno-Permic

PFV

Proto-Finno-Volgic

PFU

Proto-Finno-Ugric

PIE

Proto-Indo-European

Proto-Samoyed

Proto-Uralic

PUg

Proto-Ugric

SKES

Suomen Kielen Etymologinen Sanakirja

(Toivonen et al. 1955-1978)

Uralic

2) Broader proposals for Uralic are at best
controversial. For example, Uralic is often listed in the
literature with Altaic, in the so-called Ural-Altaic hypothesis.
This, however, is not supported, and, in fact, many of the
leading "Altaicists" have recently dismantled Altaic as a group,
showing that the evidence available does not support the
genetic affinity of several of the languages usually identified as
"Altaic" (cf. Unger 1990). Other proposed far-flung
connections which, at least for the present, lack full support are:
Uralo-Dravidian, Eskimo-Uralic, Uralo-Sumerian, etc.

3) My principal source for IE is Friedrich (1970),
who was the major stimulus for my research on this topic. I
have also consulted Pokorny (1959) and Watkins (1985).

MOTHER TONGUE

Issue 22, May 1994

Unless otherwise stated, citations and discussions of IE matters
follow Friedrich. My principal sources for Uralic (U) and
Finno-Ugric (FU) are: Toivonen et al.'s (1955-1978)
etymological dictionary of the Finnish language, henceforth
SKES; Collinder (1960, 1965); Hajdu (1965, 1969, 1975a);
Janhunen (1977); Redei (1986-1988); and Sammallahti (1988);
plus others cited less frequently. Examples are as found in the
sources, using their orthographic symbols. Although these are
not always consistent with one another, they are sufficiently
clear for the purposes of this paper.

4) According to current thinking about dating Uralic
language splits, it is believed that Proto-Uralic separated into
Proto-Finno-Ugric and Proto-Samoyed ca. 6,000 - 4,000 B.C.
Proto-Finno-Ugric split into Finno-Permian and Ugric ca.
4,000 - 2,000 B.C. (T. Itkonen 1984:350 holds ca. 2,500 B.C.
to be the latest date before the end of Proto-Finno-Ugric unity).
There is no reliable dating for the end of the Finno-Permian
period, but Proto-Finno-Volgic is calculated to have split at the
latest ca. 1,500 B.C., and Early Finno-Baltic at the latest ca.
1,000 B.C. (Korhonen 1984:66; T. Itkonen 1984:350).

5) The honeybee was unknown in Siberia, Turkestan,
Central Asia, Mongolia, and most of the rest of Asia, but was
found in eastern Europe west of the Urals with a northern
border coinciding with that of the oak.

Another diffused word that has given rise to much
speculation is "metal", with cognates in nearly all Uralic
languages meaning "copper, iron, ore, or metal" (e.g., Finnish
vaski "copper, bronze, brass"; Mordvin Viskfe "metal, wire";
Cheremis Pa "ore, metal"; Zyrian -is "metal"; Votyak -ves
"metal"; Ostyak wax "metal, iron"; Hungarian vas "iron";
Yurak wiessg "iron, money"; Yenisei bese "iron"). Since
Uralic dates to the Stone Age, such an ancient term for metal is
significant. Possibly, this stone-age population came to know
copper in the Urals, where it is found on the earth's surface
(Korhonen 1984; Hajdu 1975a:35). Nevertheless, it is
important to keep in mind that a metal term of similar shape is
found widely also in Indo-European languages, as well as in
Sumerian, so that it may be a very old Wanderwort (Joki
1973:339-340).

6) The widely-held notion that the homeland was in
the region of the Middle Volga stems from Aminoff (1873).
On the other hand, according to another also widely-held view,
the homeland would have been further east and perhaps further
north, between the Urals and the Volga-Kama-Petsora area or
on both sides of the Ural Mountains (Paasonen 1923; Zsirai
1937; Sebestyen 1952; Hajdu 1969, 1975a, 1975b).
Supporters of a third view believe that the Proto-Finno-Ugric
population, at least in its final phases, may have occupied a
rather wide area from the Urals to the Baltic, based on the
notion that hunting and fishing groups need wide territories
(Ojansuu 1907; E. Itkonen 1966; cf. Sammallahti [1984:153]
and Hajdu [1975a:37]).

In an argument from negative information, it has been
suggested that a terminus ante quern for Finno-Ugric can be
established on the basis of tree names. "Oak" has a different
name in Finno-Permian languages (e.g., Finnish tammi) from
Hungarian (tolgy). Hajdii (1975a:41) takes this to mean that by
the time the "oak" arrived to the Petsora region (3,000 - 2,000

B.C.), the Finno-Permian languages had already separated from
the Ugric branch.

7) Note that references given at the end of a list of
cognates refer to the sources of the entire cognate set.
However, occasionally forms cited from some other source are
inserted into these lists, in which case the references appear
immediately adjacent to these inserted additional forms.

8) Rasanen (1955:28) reported similar forms in other
languages in the general region, e.g.:

Turkish (M.)

Chuvash

Soyot

kajrj "Scots pine"
xira "Scots pine"
xady "Scots pine"

9) Similarities from other languages of the region
include:

Tungus se-kta "Salix viminalis"

Uighur (Turkish) sogut "willow" (Rasanen 1955:27)

10) Compare the phonetically similar but distinct
PFU *pec3 "willow" (Redei 1986-1988:367):

Votyak puc'i "bud, willow catkin"

Zyrian paca

Hungarian fuz "willow"

Compare also forms from other languages:

Tungus fodo "willow"

Korean padil "willow" (Rasanen 1955:27)

11) Watkins (1985:47) presents the possibility that
the Latin form is from *pei3- "to be fat, to swell" with a
suffixed zero-grade, *pl-nu- ("pine tree, yielding a resin").

12) Sinor (1969:279) has pointed out some
similarities in other languages of the area:

Yakut tit "pine"

Altaic Turkic dialect fit "larch"

Rasanen (1955:26) compared also:

Tungus sakso-kumar "conifer"

Goldic sese "Larix dahurica"

13) Hajdu (1969:257) considered another possible
connection between IE and U in citing the conifer terms:

IE
PU

*kuse "spruce (Picea)"
**kuse "spruce (Picea)'

I do not find the IE term confirmed elsewhere, however. The
U form is based on one of the most solid of U cognate sets.

14) For additional evidence of areal diffusion,
compare Tungus morirjo "cloudberry", which Rasanen
(1955:28) thought to be perhaps borrowed from Samoyed, and

MOTHER TONGUE

Issue 22, May 1994

Tobol (Tatar) myrak "cloudberry", which Rasanen (1955:28)
believed possibly to be a loan from Vogul.

15) Conceivably one could also compare the IE forms
to another etymon:

PFV *paksna "linden"

Estonian pahn "old linden"

Mordvin pekse, pasa "linden"

Cheremis pitss, pists "linden"

(Redei 1986-1988:726)

16) Similar forms from a different cognate set have
sometimes erroneously been attributed to this set.
Nevertheless, it is interesting to note the similarity of forms
representative of PU **jakk3 "pine or spruce forest" (Redei
1986-1988:107) with the compared IE forms.

17) Watkins (1985:12) reconstructs *deru "tree,
wood", with primary senses "to be firm, solid, steadfast". He
gives a variant form *derw-, in Germanic *terw-, which has the
Old English cognate te(o)ru "resin, pitch".

18) Compare also:

Turkish

Sagai

Mongolian

REFERENCES

jumurt "bird-cherry"
namyrt "bird-cherry"
zimu-gu-sun "bird-cherry"

(Rasanen 1955:27)

Aminoff, Torsten. 1873. "Lyhyt Silmays Itaisten
Suomensukuisten Kansain Historiaan ennen heidan
Joutumistaan Wenajan Vallen alle [A Brief Look at the
History of the Eastern Peoples who are related to the
Finns before they came under Russian Domination]".
Savo-karjalaisen Osakunnan Albumi II, pp. 76-77.
Helsinki: Koitar.

Benveniste, Emile. 1955. "Homophones radicales en indo-
europeen". Bulletin de la Societe de Linguistique de
Paris 5XAA-29.

Collinder, Bjorn. 1934. Indo-uralisches Sprachgut: Die
Urverwandtschaft zwischen der indoeuropaischen und
der uralischen (finnisch-ugrisch-samojedischen)
Sprachfamilie. (= Uppsala Universitets Arsskrift, 1934.)
Uppsala: A. B. Lundequistska.

Collinder, Bjorn. 1954. "Zur indo-uralischen Frage". Acta
Universitatis Upsaliensis (= Sprdlvetenskapliga
sdllskapets i Uppsala forhandlingar, 1952-1954) 10.19-
91. Uppsala: Amqvist & Wiksells.

Collinder, Bjorn. 1955. Fenno-Ugric Vocabulary: An
Etymological Dictionary of the Uralic Languages.
Stockholm: Almqvist & Wiksells.

Collinder, Bjorn. 1960. Comparative Grammar of the Uralic
Languages. Stockholm: Almqvist & Wiksells.

Collinder, Bjorn. 1965. "Hat das Uralische Verwandte?: Eine
sprachvergleichende Untersuchung". Acta Universitatis
Upsaliensis 1:4.109-180.

Decsy, Gyula. 1988. "Slawischer EinfluB auf die uralischen
Sprachen", in Sinor 1988, pp. 613-637.

Decsy, Gyula. 1990. The Uralic Protolanguage: A
Comprehensive Reconstruction. Bloomington, IN:
Eurolingua.

Dolgopolsky, Aaron B. 1984. "On Personal Pronouns in the
Nostratic Languages". Linguistica et Philologica:
Gedenkschrift fur Bjorn Collinder (1894-1983), ed. by
Otto Gschwantler, Karoly Redei, and Hermann Reichert,
pp. 65-112. Vienna: Wilhelm Braumiiller.

Friedrich, Paul. 1970. Proto-Indo-European Trees: The
Arboreal System of a Prehistoric People. Chicago, IL:
University of Chicago Press.

Gallen, Jarl, ed. 1984. Suomen Vdeston Esihistorialliset Juuret
[Prehistoric Roots of the Finnish People]. (= Bidrag till
Kannedom av Finlands Natur och Folk, Utgivna av
Finska Vetenskaps-Societeten, 131.) Helsinki: Societas
Scientiarum Fennica.

Greenberg, Joseph H. 1987. Language in the Americas.
Stanford, CA: Stanford University Press.

Greenberg, Joseph H. Forthcoming. Indo-European and Its
Closest Relatives: The Eurasiatic Language Family.
Stanford, CA: Stanford University Press.

Haarmann, Harald. 1976. Aspekte der Arealtypologie: Die
Problematik der europdischen Sprachbilnde. Tubingen:
Gunter Narr.

Hajdu, Peter. 1964. "Uber die alten Siedlungsraume der
uralischen Sprachfamilie". Acta Linguistica Academiae
Scientiarum Hungaricae 14.47-82. Budapest.

Hajdu, Peter. 1969. "Finnougrische Urheimatforschung".
Ural-altaische Jahrbiicher 41 .252-264.

Hajdu, Peter. 1975a. "Sukulaisuuden Kielellisia Taustaa [The
Linguistic Background of Genetic Relationship]".
Suomalais-ugrilaiset, ed. by Peter Hajdu, pp. 11-51.
Pieksamaki: Suomalaisen Kirjallisuuden Seura.

Hajdii, Peter. 1975b. Finno-Ugrian Languages and Peoples.
London: Andre Deutsch.

Hakulinen, Lauri. 1968. Suomen Kielen Rakenne ja Kehitys
[The Structure and Development of the Finnish
Language]. Helsinki: Otava.

Itkonen, Erkki. 1966. Suomalais-ugrilaisen Kielen- ja
Historiantutkimuksen Alalta [From the Field of Finno-
Ugric Linguistic and Historical Investigation] . (=
Tietolipas, 20.) Helsinki: Suomalaisen Kirjallisuuden
Seura.

Itkonen, Terho. 1984. "Suomessa Puhutun Suomen
Kantasuomalaiset Juuret [The Proto-Finnic Roots of
Finnish Spoken in Finland]", in: Gallen 1984, pp. 347-
363.

Jakobson, Roman. 1938. "Sur la theorie des affinites
phonologiques entre les langues". Actes du Quatrieme
Congres International de Linguistes, pp. 48-58.
Copenhagen: Munksgaard. (Reprinted as an appendix to
Principes de Phonologie by N. S. Troubetzkoy, pp. 351-
365. Paris: Editions Klincksieck, 1949. — Translated in
English as "On the Theory of Phonological Associations
among Languages", in Allan R. Keiler, ed.: A Reader in
Historical and Comparative Linguistics, pp. 241-252.
New York, NY: Holt, Rinehart & Winston, 1972.)

Janhunen, Juha. 1977. Samojedischer Wortschatz:
Gemeinsamojedischer Etymologien. (= Castrenianumin

MOTHER TONGUE

Issue 22, May 1994

Toimitteita, 17 .) Helsinki: Helsingin Yliopisto.

Joki, Aulis J. 1963. "Omenan Vaellus [The Wandering of the
Apple]". Virittdjd 1 963 .134-141. [German translation as
Joki 1964.]

Joki, Aulis J. 1964. "Der wandernde Apfel". Studia
Orientalia 28/12.7-12.

Joki, Aulis J. 1973. Uralier und Indogermanen. (=
Suomalais-ugrilaisen Seuran Toimituksia / Memoires de
la Societe Finno-ougrienne, 151.) Helsinki: Suomalais-
Ugrilainen Seura.

Joki, Aulis J. 1974. Class handouts from course, "Suomalais-
ugrilaisen Kultuuri [Finno-Ugric Culture]". Unpublished
manuscript, University of Helsinki.

Kaiser, Mark and Vitalij Shevoroshkin. 1988. "Nostratic".
Annual Review of Anthropology. 17.309-329.

Katz, Hartmut. 1975. Generative Phonologie und
phonologische Sprachbunde des Ostjakischen und
Samojedischen. Munich: WilhelmFink.

Koivulehto, Jorma. 1976. "Vanhimmista Germaanisista
Lainakosketuksista ja niiden Ikaamisesta, I ja II [On the
Oldest Germanic Loan Contacts and their Dating, I and
II]". Virittdjd 80.33-45, 257-281.

Koivulehto, Jorma. 1980. "Die Datierung der altesten
germanischen Lehnworter im Finnischen". Congressus
Quintus Internationalis Fenno-Ugristarum, part 7, ed. by
Osmo Ikola, pp. 73-78. Turku: Suomen Kielen Seura.

Koivulehto, Jorma. 1984. "Itamerensuomalais-germaaniset
Kosketukset [Contacts between Germanic and Finnic]", in
Gallen 1984, pp. 191-205.

Koppen, Friedrich Theodor. 1886. Materialy k voprosu
PervonatsaVnoj Rodine i Pervobytnom Rodstve Indo-
jeropejskogo i Finno-ugorskogo plemeni. St. Petersburg.

Koppen, Friedrich Theodor. 1890. "Ein neue

tiergeographischer Beitrag zur Frage iiber die Urheimat
der Indoeuropaer und Ugrofmnen". Das Ausland, pp.
1001-1007. Stuttgart.

Korenchy, Eva. 1988. "Iranischer EinfluB auf die uralischen
Sprachen", in Sinor 1988, pp. 665-681.

Korhonen, Mikko. 1984. "SuomalaistenSuomalais-ugrilainen
Tausta Historiallis-vertailevan Valossa [The Finno-Ugric
Background of the Finns in Light of Comparative-
Historical Linguistics]", in Gallen 1984, pp. 55-71.

Lehtiranta, Juhani. 1989. Yhteissaamelainen Sanasto
[Common Lapp Vocabulary]. (= Suomalais-ugrilaisen
Seuran Toimituksia / Memoires de la Societe Finno-
ougrienne, 200.) Helsinki: Suomalais-Ugrilainen Seura.

Ojansuu, Heikki. 1907. "Suomen Suvun Esihistoria [The
Prehistory of the Finnish People]". OmaMaal. Porvoo.

Paasonen, Heikki. 1923. Beitrdge zur Aufliellung der Frage
nach der Urheimat der finnisch-ugrischen Volker. (=
Turun Suomalaisen Yliopiston Julkaisuja, 1:5.) Turku:
Turun Suomalaisen Yliopisto.

Pokorny, Julius. 1959. Indogermanisches etymologisches
Worterbuch. Munchen: A. Francke.

Rasanen, Martti. 1955. "Uralaltaische Wortforschung".
Studia Orientalia, Editit Societas Orientalis Fennica
18/3.1-59. Helsinki.

Redei, Karoly, ed. 1986-1988. Uralisches etymologisches
Worterbuch. Band I, Lieferungen 1-5, Budapest:

Akademiai Kiado. Band II, Lieferungen 6-7, Wiesbaden:

Otto Harrassowitz.
Redei, Karoly. 1988. "Die altesten indogermanischen

Lehnworter der uralischen Sprachen", in Sinor 1988, pp.

638-664.
Ruhlen, Merritt. 1987. A Guide to the World's Languages.

Vol. 1: Classification. Stanford, CA: Stanford

University Press.
Ruhlen, Merritt. 1990. "An Overview of Genetic

Classification". The Evolution of Human Languages. (=

SFI Studies in the Sciences of Complexity, Proceedings).

Ed. by John A. Hawkins & M. Gell-Mann, vol. 10, pp.

1-22. Reading, MA: Addison- Wesley.
Sammallahti, Pekka. 1984. "Saamelaisten Esihistoriallinen

Tausta Kielitieteen Valossa [The Prehistoric Background

of the Lapps in the Light of Linguistics]", in Gallen 1984,

pp. 137-156.
Sammallahti, Pekka. 1988. "Historical Phonology of the

Uralic Languages", in Sinor 1988, pp. 478-554.
Sebeok, Thomas S. 1950. "The Importance of Areal

Linguistics in Uralic Studies". Memoires de la Societe

Finno-ougrienne 98.99-106. Helsinki.
Sebestyen, Irene. 1952. "Zur Frage des alten Wohngebiets der

uralischen Volker". Acta Linguistica Academiae

Scientarum Hungaricae 1.330-342. Budapest.
Shevoroshkin, Vitalij. 1989. "Methods of Interphyletic

Comparisons". Ural- A Itaische Jahrbilcher 61.1 -26 .
Sinor, Denis. 1969. "Geschichtliche Hypothesen und

Sprachwissenschaft in der ungarischen, finnisch-
ugrischen und uralischen Urgeschichtsforschung". Ural-

Altaische Jahrbilcher 41 .273-28 1 .
Sinor, Denis, ed. 1988. The Uralic Languages: Description,

History, and Foreign Influences. Leiden: E. J. Brill.
SKES — see Toivonen et al. 1955-1978.
Suhonen, Seppo. 1988. "Die baltischen Lehnworter der

finnisch-ugrischen Sprachen", in Sinor 1988, pp. 596-

615.
Tauli, Valter. 1955. "On Foreign Contacts of the Uralic

Languages." Ural-Altaische Jahrbilcher 27.7-3 1 .
Thomason, Sarah Grey & Terrence Kaufman. 1988.

Language Contact, Creolization, and Genetic Linguistics.

Berkeley, CA: University of California Press.
Toivonen, Y. H. 1953. "Suomalais-ugrilaisesta Alkukodista

[On the Finno-Ugric Homeland]". Virittdjd 1953.5-35.
Toivonen, Yrjo H., Aulis J. Joki, Erkki Itkonen, & Reino

Peltola. 1955-1978. Suomen Kielen Etymologinen

Sanakirja [Etymological Dictionary of the Finnish

Language]. Vol. I, 1955, Toivonen; vol. II, 1959,

Toivonen, Itkonen, Joki; vol. Ill, 1962; vol. IV, 1969,

Itkonen, Joki; vol. V, 1975; vol. VI, 1978, Itkonen, Joki,

Peltola. Helsinki: Suomalais-Ugrilainen Seura.
Uesson, Ants-Michael. 1970. On Linguistic Affinity: The

Indo-Uralic Problem. Malmo: Estonian Post Publishing.
Unger, J. Marshall. 1990. "Summary Report of the Altaic

Panel". Linguistic Change and Reconstruction

Methodology, ed. by Philip Baldi, pp. 479-482. Berlin:

Mouton de Gruyter.
Watkins, Calvert. 1985. The American Heritage Dictionary of

Indo-European Roots. Boston, MA: Houghton Mifflin.

MOTHER TONGUE

Issue 22, May 1994

Zeps, Valdis. 1962. Latvian and Finnic Linguistic
Convergences. (= Uralic and Altaic Series, 9.)
Bloomington, IN: Indiana University Press.

Zsirai, Miklos. 1937. Finnugor Rokonsdgunk. Budapest.

The following report is reprinted with permission from the
FAIES Newsletter, vol. Ill (February 1994), no. 2, pp. 6-7.

REPORT FROM THE FIELD:
THE TOCHARIANS?

KARLENE JONES-BLEY
UCLA

This past summer while in Siberia, I had the
opportunity to talk to Sergei Grigorevich Skobelev, chief
archaeologist of the Laboratory of Humanitarian Research,
Novosibirsk State University. He works primarily in the period
of the Iron Age and is particularly interested in the Tashchtik
culture (2nd c. BC - 5th c. AD) and its predecessor, the Tagar
culture (7th c. - 2nd c. BC). These cultures existed along the
southern border of Western Siberia neighboring Northern
Kazakhstan. The central area of the cultures extended from
Krasnoyarsk City south to Abakan on the Yenisei River and
along it to Sayanogorsk. Dr. Skobelev's work touches on the
archaeological identification of the Tocharians, an especially
difficult problem. Below are a few notes from two
conversations I had with him which may be of interest to those
interested in the cultural aspects of the Tocharians. (Also, see J.
P. Mallory, In Search of the Indo-Europeans, 1989.)

The last period of the Tagar culture is related to the
Tashchtik culture, and occupied an area in the southern part of
the later Tashchtik culture along the Yenisei above
Sayanogorsk. Skobelev is quite convinced that these Tashchtik
people are the Tocharians and bases this conclusion on the
archaeological data he has discovered.

At the end of the Tagar and beginning of the
Tashchtik period, the kurgans, which mark the graves of these
people, are larger than those of the preceding Bronze Age.
They frequently contain many large stones on the inside and
out, and sometimes multiple burials. One such kurgan with 200
skeletons is a mass grave with people "just tossed in." The
Tagar people have many similarities to the Scythians, while the
Tashchtik are physically a mixture of Scythian types and
Mongols. In the Tashchtik graves, there are many gypsum
death masks — almost all of men — both Mongolian and
European types but mostly Mongolian. At the end of the
Tashchtik period, men were frequently cremated but the
women who were with them were not. This may relate to
different cultural backgrounds for the men and women or
perhaps just a gender difference in burial ritual, in light of the
fact that women are buried with poor goods. The usual burial is
one man with 4 sacrificed women . Typically, there was a hole
in the back of the female's skull where the brain had been
removed. Skobelev and his colleagues are unable to tell if the
hole was the cause of death or occurred immediately after

death. The women were European and the men Mongoloid. A
similar situation has been found in Pazyryk graves which are
found just south of the Tashchtik area. (For a full account of the
Pazyryk graves, see Rudenko, Sergei L., Frozen Tombs of
Siberia: The Pazyryk Burials of Iron-Age Horsemen, 1970.)
Only a few women were buried with great wealth and these
also had masks. There is clear influence from China because of
the presence of ivory hair pins.

The Huns, who were Turkic speakers, followed the
Tashchtik in this area, and it was the Huns who brought iron to
these people. Hun graves can be distinguished from Tashchtik
on several accounts. For example, many Hun graves had horses
but none are found in Tashchtik or Tagar graves. Iron horse bits
came with Huns. Bows and arrows are found in female as well
as male Hun graves. Women skeletons also exhibit battle
wounds, perhaps confirming the statements of the 4th c. Roman
historian Ammianus Marcellinus, who wrote about Hun
women in battle with Germans. Hun men wore earrings only in
left ear. Stelae have often been found on kurgans similar to
those found on IE kurgans — some carved, some not. There
can be several to a single grave. Pictures of people, animals,
horses, axes, houses, and daily life appear on both the stelae as
well as on the stone walls of the graves. Huns mixed with
Europeans, but they were mostly Mongoloid.

It may, therefore, be that the Huns drove out the
Tashchtik people south to the Tarim basin. This basin is located
directly south of the area mentioned above and it is the Tarim
basin which is the historical location of the Tocharians. The
question is: how far back can we take this? Can these people,
the Tashchtik, be linked to the Afanasievo culture, which dates
back to the third millennium BC and whose remains have
similarities to the Yamnaya culture of the Pontic-Caspian area?

For those who may be interested in this problem,
contact Sergei Grigorevich Skobelev, Chief Archaeologist of
the Laboratory of Humanitarian Research, Novosibirsk State
University, Pirogova, 2, 630090 Novosibirsk-90, Russia.

ON THE GENETIC CLASSIFICATION
OF BASQUE

JOHN D. BENGTSON

Minneapolis, MN

During the past few decades, the topic of the genetic
classification of the Basque language, when not avoided
entirely, has divided scholars into two main camps: those
favoring (1) the Afroasiatic origin or (2) the Caucasian (or
Dene-Caucasian) origin of Basque. To the first school belong
Hugo Schuchart, G. von der Gabelentz, and, until recently, the
late Hans Mukarovsky. To the second belong C. C. Uhlenbeck,
Alfredo Trombetti, Rene Lafon, Karl Bouda, Georges Dumezil,
and, more recently, Vjaceslav Cirikba, Vaclav Blazek, and the
present writer. In a recent issue of this journal, it was reported
that A. B. Dolgopolsky had joined the Basque-Afroasiatic
school, and it was stated that "the Basque-Berber connection is

MOTHER TONGUE

Issue 22, May 1994

a serious competitor to the proposed link of Basque to
Caucasic" (McCall and Fleming 1994:26). With respected
scholars and dozens of lexical comparisons on both sides, how
do we decide?

I propose that we carefully examine the evidence for
both proposals in terms of the two major criteria for genetic
classification: (1) similarity in core (basic) lexemes and (2)
similarity in grammatical formatives. For criterion (1), it is
necessary to focus on the core (or basic) lexicon of the
languages involved: the semantic fields least likely to erode
over millennia of time. For that purpose, the comparisons
listed here fall within Dolgopolsky's (1964) hierarchy of the
twenty-five most stable meanings.

1 . (a) Basque ni "I" (ergative nik), compared by Mukarovsky
with Berber ndk, ndkki "I", Egyptian ink, Akkadian anaku,
etc.; (b) Basque ni "I": cf. Caucasian *ni (Lak na, Dargi
nu id.).

2. (a) Basque gu "we", compared by Mukarovsky with
Berber ndkni "we"; (b) Basque gu "we": cf. Caucasian
*l$[u] "we" (inclusive: Avar n-S, Andi Si, Agul xi-n,
Tabasaran uxu).

3. Basque bi "two": cf. Caucasian *q'wi (Ubykh t'-q' w a,
Archi q' w e 9 Udipa). (Basque bi from *G w i 9 influenced by
bat "one", cf. English four, five; Latin quattuor, quinque).

4. (a) Basque hi "thou" (2nd person singular), compared by
Mukarovsky with Berber kiyyi "thou"; (b) Basque hi
"thou": cf. Caucasian *gu (Taskhur yu, Dargi hu,
Chechen ho, ah)', Burushaski gu- "thou".

5. Basque zu [su] "you" (2nd person plural): cf. Caucasian
*zwV (Lak zu, Chechen su, Kabardian^e, etc.).

6. Basque ze-r "what?", ze-in "which?": cf. Caucasian *sa
(Ubykh sa "what?", Dargi se, etc.); Burushaski be-s, be-
sa-n "what?".

7. Basque no-r "who?", no-iz "when?": cf. Caucasian *nV
(Agul na, Lezgi ni "who?", Tsezi ne-ti "when?");
Burushaski (Werchikwar) ana, dne "where, whither?".

8. (a) Basque mihi "tongue", compared by Mukarovsky with
Semitic: Amharic mdla's "tongue", Argobba mala's;
Chadic: Boghom rjdlis; (b) Basque mihi "tongue": cf.
Caucasian *melc 7 "tongue" (Circassian bzd, Andi mic 7,
Tabasaran melz, etc.); Burushaski melc "jaw" (metonymy).

9. (a) Basque i-zen, (Vizcayan) u-zen "name", compared by
Mukarovsky with Berber isem, isdm, (Zenaga) isem,
Arabic 'ism- (and other cognates in Semitic, Cushitic,
Chadic, Omotic); (b) Basque i-zen, (Vizcayan) u-zen
"name": cf. Burushaski sen-as "to say, to tell, to call, to
name", sen-as "named". The Basque alternation of
prefixes suggests that the word may be archaic rather than
borrowed. Cf, more remotely, Na-Dene: Tlingit sa,
"name, voice", -sa, -sen "to name, to breathe".

10. (a) Basque begi "eye", compared by Mukarovsky with
Egyptian b>k "to see", and forms like the following in

Cushitic: Oromo be'kw "to know"; Omotic: Bencho
bek'a id.; (b) Basque begi from *beryi "eye": cf.

Caucasian *?wil£i "eye" (Tabasaran ul, Hunzib hare,
Khinalug pil, Avar ber, Chechen Marg, etc.); Burushaski
-l-ci(n) "eye (of needle)", -vl- "eye" (in compounds).

11. Basque bihotz (bi-hotz) "heart": cf. Burushaski -as
"heart" (from *has). (For the Basque fossilized class
prefix be-/bi-, cf. also §§15, 32, below.)

12. (a) Basque hortz "tooth", compared by Mukarovsky with
Berber (Rif) a-qarrus "tooth"; (b) Basque hortz "tooth":
cf, by metathesis *hilcV ~ Caucasian *cilhV "tooth"
(Circassian ca, Dargi cula, Chechen cer-g, etc.);
Burushaski -a-was (-me), (Werchikwar) -base (-me)
"molar (tooth)"; (c) Basque hortz "tooth": cf Caucasian
*gwa/jwe "fang, canine tooth" (Lak karci "tooth", Avar

gozo "fang, canine, etc.").

13. Basque ez, (Vizcayan) ze "not": cf. Caucasian *cd/*6'9
"not" (Chechen ca, -c, Ingush cd-, Batsbi co-); Burushaski
(Werchikwar) acho "not yet".

14. Basque oskol (Vizcayan) "nail, claw", compared by
Mukarovsky with Berber isker "nail". (The Basque word
has many dialectal variants: ezkazal, azazkal, azuskulii,
etc., some with the meaning "hoof).

15. Basque be-hatz (Labourdin) "nail": cf. Caucasian *kwac 'e
"paw" (Avar k w ac\ Dargi kac'a, etc.).

16. Basque nigar, negar "tears, weeping": cf. Caucasian
*newq 'u (Dargi nery "tear", Lak maq' id., Chechen

not'q'a "pus"); Burushaski nagei ~ magei "boil" (sore
with pus).

17. (a) Basque mama "water, potable liquid" (child speech),
compared by Mukarovsky with Berber ama-n "water",
Arabic ma-'-un, etc.; (b) Basque mama "water, potable
liquid": cf. Burushaski mamu "milk, sap (of crops)".

18. Basque hur "water" (usual word): cf. Caucasian *hwiri
"lake" (Avar hor, Lezgi wir ~ ur, etc.); Burushaski hur-
o-yo "sweat", hur-iginas "stream, current, wave".

19. Basque u-hin (Labourdin) "wave (in the water)": cf.
Caucasian *xanfii "water" (Chechen Xi, Avar Urn, Dargi

*xin, etc.).

20. (a) Basque hil "to die, to kill; dead", herio "death",
compared by Mukarovsky with Semitic h r g "to kill",
Cushitic: Khamir qirw, etc.; (b) Basque hil "to die, to kill;
dead": cf. Caucasian *-i(w)K 9 V "to die, to kill" (Circassian
K'a-n, Chechen -al- id., Hunzib -iK'- "to kill").

21. Basque bartz "nit": cf. Caucasian *nemje "louse"
(Abkhaz a-c'a, Lak nac\ Chechen meza, etc.).

22. (a) Basque hil- "moon", compared by Mukarovsky with
Berber (Tuareg) ta-lli-t "new moon"; Arabic hill-un
"appearance of the new moon (Monatsanfang)"; (b)
Basque hil-, hil-argi, il-az-ki (Basse Navarre) "moon"; cf.
Burushaski halanc ~ halanz "moon".

23. Basque hil-argi, argi-zagi "moon"; cf. Caucasian *(wi-)
rdq 'A "sun" (Ubykh n-dya, Lezgi ray, Lak bary, etc.).

24. Basque esku "hand": cf. Burushaski hisk "wrist, back of
the hand", (Werchikwar) hesk "base of the palm of the

MOTHER TONGUE

Issue 22, May 1994

hand and thumb".

25. (a) Basque bortz, host "five" (from "hand"), compared by
Mukarovsky with Berber a-fus "hand"; (b) Basque bortz,
bost "five" (from "hand"), bost-eko (Guipuzcoan,
Vizcayan) "either one of the hands (cualquiera de las
manos)": cf. Caucasian *borcV (Tsezi besi "fist",
Tabasaran bac "paw", Avar purci "ham, [animal's]
gammon"); Burushaski bac-in "thigh".

26. Basque u-khare "wrist" (u-khal- in compounds): cf.
Caucasian *q w[a]l_?V (Ubykh q a-p 'a "hand", Avar q ' w al

"arm, sleeve", Khwarshi q 'e "shoulder").

27. Basque a-hur "palm (of the hand)": cf. Caucasian *kwll?i
"hand" (Archi, Udi kul, Avar k w er, Hunzib koro, etc.).
Basque h ~ Caucasian *k(w) is regular (cf. §15, above).

28. Basque gau "night" (gab- in compounds): cf. Caucasian:
Lezgian *Xam (Taskhur Xam "night", Tabasaran Xab-as

"evening"); Abkhaz -Xa "night".

29. (a) Basque a-dar "branch, horn", compared by
Mukarovsky with Berber (Tuareg) dele "branch (Zweig)",
cf. also Coptic tar "point, end, branch"; (b) Basque a-dar
"branch, horn": cf. Caucasian *%wirV (Avar Kar "horn",
Chechen kur "[wild animal's] horn"); Burushaski tur
(unprefixed), -Itur (with prefix) "horn".

30. Basque bethe "full": cf. Caucasian *bVtV (Kabardian
bdda "strong, solid", Chechen butu "solid, dense", etc.);
Burushaski bwt "much, many, numerous, very".

3 1 . Basque oso "whole, entire, healthy": cf. Caucasian *fioc 'V
"full, to fill" (Ubykh -za, Lak -u-c 'w, Tabasaran ac 'u, etc.).

32. Basque be-larri "ear" (Vizcayan, Guipuzcoan): cf.
Caucasian *[eH[e "ear" (Abkhaz -ld-mha, Dargi lihi,

Chechen ler-g). Basque be-harri, be-arri in other dialects
probably was influenced by the verb beha- "to listen, to
look".

DISCUSSION OF LEXICAL COMPARISONS

Taking a raw "score" on the comparisons, without any
evaluation, we have 13 Basque- Afroasiatic comparisons, 26
Basque-Caucasian comparisons, and 17 Basque-Burushaski
comparisons. On the face of it, all these scores could strongly
suggest genetic affinity of the languages in question, but the
narrow range of scores (13 ~ 26) would not be considered
decisive by many linguists. Let us take a closer look at the
comparisons:

(la) Basque ni ~ Berber ndk(ki): Basque ergative
ni-k, but -k is a distinct morpheme that is suffixed to nouns as
well, so the comparison is really ni with Afroasiatic *? n k or
the like (derived by Blazek from *?an-aku). The Caucasian
*ni is much more likely.

(2a) Basque gu ~ Berber ndkni: Once again,
comparison with wider Afroasiatic (e.g., Semitic *nahna/u)
indicates two nasals, neither of which is found in Basque.

Again, comparison with Caucasian *fc[ii] is more
straightforward. *t$ is a lateral fricative, and the close
relationship between laterals and velars is abundantly evident
both in Basque and Caucasian languages: e.g., Caucasian
*roA'f "meat, flesh" (with globalized lateral affricate = ti')

becomes Andi riK'i, but Lak dik\ Khinalug lika, with velars
like Basque a-ragi "meat, flesh".

(8a) Basque mihi ~ Amharic mdla's, etc.: wider
comparison in Afroasiatic shows that the root is really *lis, and
the m- is apparently a prefix (or, the words come from the root
*//?s"tolick).

(8b) The phonetic development in Basque may have
been *milci > *misi > *mixi > mihi. The first stage, with
retroflex -/c-, is preserved in Burushaski. The development of
s or retroflex s to a velar fricative (jc, x w ) is widely attested (e.g.,
in Castilian Spanish, Swedish, Slavic). Finally, there is good
reason to believe that earlier *x is one of the sources of modern
Basque h (cf. §33, below). A parallel case could be Basque
behi "cow" < *bexi < *besi < *berci, cf. Caucasian *bharc 'wV

(Avar boc'i "cattle").

(10) Basque begi- Egyptian blk, etc.: the meaning is

verbal rather than anatomical. The Afroasiatic words do not
account for the r in Souletin ber-, while the Dene-Caucasian
explanation does.

(12) Basque hortz ~ Berber a-qarrus "tooth": If
affinity with broader Afroasiatic is sought, the usual word for
"tooth" (*s n) is quite different. I prefer the solution of (12b):
the form *hilcV (eliminated in Caucasian proper by the
metathetic *cilhV) corresponds perfectly with Basque hortz
and Burushaski *-h w as- (see the correspondences established
by Bengtson 1991c).

(17, 18) The form mama "water" is associated with
infantile language (and universal Lallworter), while the usual
word hur is common Dene-Caucasian (cf. also Yeniseian *xur
"water").

(22, 25, 29) Again, the Dene-Caucasian explanations
are semantically and/or phonetically more straightforward than
the Afroasiatic. Note the precise morphological corres-
pondence of the Basse Navarre and Burushaski words for
"moon": il-az-ki ~ hal-anc (22b).

With these evaluations taken into account, the case for
genetic affinity of Basque with Caucasian and Burushaski
becomes much more decisive. Barely a handful of Basque-
Afroasiatic comparisons now seem plausible, while the
Basque-Caucasian and Basque-Burushaski parallels are both
more numerous and more plausible, semantically and
phonetically, than the Afroasiatic alternatives. Most important,
Schuchardt, Mukarovsky, and other scholars have found no

MOTHER TONGUE

Issue 22, May 1994

parallels between Basque and Afroasiatic interrogatives and
negatives, though parallels in these fields are found between
Basque, Caucasian, and Burushaski (§§6, 7, 13). I conclude
from this that Basque is closest to Caucasian and Burushaski,
while Afroasiatic, if related at all, must be more distant. This
is the same conclusion Trombetti, Lafon, and others arrived at
several decades ago.

There remain, however, three fairly acceptable
Basque- Afroasiatic comparisons in the above list: (4a) Basque
hi "thou" ~ Berber kiyyi, (9a) Basque izen "name" ~ Berber
issm, and (14) Basque oskol "nail" ~ Berber isker. How are
these to be explained, in the face of a likely Dene-Caucasian
affiliation of Basque? In my opinion, (4a) could represent
evidence of a deeper relationship between Dene-Caucasian and
Afroasiatic (cf. Starostin 1989), while (9a) and (14) more likely
represent loanwords (if we discount comparison [9b]) by
Basque from Berber or a Berber-like language. Centuries of
contact with the Tartessians, on the far southwest corner of
Iberia (Spain-Portugal), may have been the source of such
loanwords (cf. McCall and Fleming 1994:25).

A notable fact about Basque-Berber parallels is that,
while the presumed genetic cognates are phonetically and/or
semantically difficult, cultural or technological words are very
similar, viz., Basque bide "road" ~ Berber a-brid, ta-brida;
Basque izten "awl" ~ Berber t-isten-t (t is the feminine marker);
Basque matel "cheek" ~ Berber a-mwddl (words denoting parts

of the face, unlike most other anatomical words, are not
uncommonly borrowed, e.g., English face, jaw from French).
This tells us that most of the similarities between Basque and
Berber are due to historical contact rather than genetic origin.

There are in fact several layers in the Basque lexicon.
To give just a few examples: Kartvelian (Basque khako, krako
"hook" ~ Georgian k'akvi id.; Basque larre "pastureland" ~
Svan lave "meadow"; Basque apho ~ aphu "toad" - Svan apXw
"frog"); Egyptian (Basque zazpi "seven" - Coptic sasf, sasf
"seven"; Basque berri "new" ~ Coptic beri, brre "new,
young"; Basque arrain "fish" - Coptic rame, raame, rami
"fish"); Berber (as noted above; cf. also Basque zilh-ar "silver"
- Basque a-zerf, Basque urhe "gold" ~ Berber uriy "yellow" ~

or Laz orXo, Georgian oXro "gold ); and finally Latin
(Basque luma "feather" ~ Latin pluma "feather"; Basque bilho
"hair" - Latin pilu- "hair" are among the few Latin words that
have penetrated into Basque). But all this has affected the basic
vocabulary of Basque very little. The great majority of
meanings on the Swadesh 100-word list can be accounted for
by Caucasian or Dene-Caucasian. Just a few examples will be
given here:

33. Basque txori [cori] ~ xori [sori] "bird": cf. Caucasian
*c'wilV (Avar c'orolo "quail", Chamalal c'or "bird", Lak
c'il-mu "small bird"); Burushaski cili (babwk) "a very
small bird"; Sino-Tibetan: Tibetan m-chil-pa "sparrow"
(Blazek and Bengtson 1995, §58).

34. Basque hotz "cold": cf. Caucasian: Avar k w ac- "coolness,
frost"; Yeniseian: Pumpokol kvc-idin "cold"; Na-Dene:
Athabaskan *k 'az? "(to be) cold" (Hupa k 'ac\ Navajo

-k'dz, -k'as) (Blazek and Bengtson 1995, §185).

35. Basque su "fire" ~ Caucasian *c 'aji "fire" (Avar, Dargi
c 'a, Lak c 'u, Chechen c 'e, etc.); Burushaski si "fireplace",
su-tin "hearthstones", (Werchikwar) su-tum "hearth,
stone" (Bengtson 1990, §25).

36. Proto-Basque *a-x w o "mouth" (aho, ago, abo, ao in the

various dialects): cf. Sino-Tibetan *Kho(w)H "mouth"
(Dimasa khu, Old Chinese *khu); Yeniseian *Xowe
"mouth" (Ket qo\ Assan xoboj); Na-Dene *Xw? "tooth"

(EyakXw-, Kutchin (d)-ywo) (Blazek and Bengtson 1995,

§5).

37. Basque i-zar "star": cf. Caucasian *Ha-jwar?i "star":

Abkhaz -jac ' w a, Akhwakh c ' w ari, Avar c' w a, etc. (Cirikba
1985:102, §84).

Let us now go on to the other criterion of genetic
classification, which is grammatical affinity. As we have
already seen, pronominal parallels between Basque and
Afroasiatic are scanty at best, and agreements between the
respective interrogatives and negatives are nonexistent. Such
common Afroasiatic features as the t feminine and internal a-
plurals are totally unknown in Basque.

But positive evidence, not negative, is decisive.
Numerous grammatical parallels between and among Basque,
Caucasian, and Burushaski have been detailed by Trombetti,
Lafon, Dumezil, and others (see, e.g., Cirikba 1985; Bengtson
1990, 199 Id, 1991e, 199 If, 1993). Among these are the
fossilized noun prefixes a-, e-/i-, o-/u-, be-/bi- of Basque,
which have been compared with the class systems of Caucasian
and Burushaski. (Examples of the prefixes can be found in this
study: §§9, 11, 15, 19, 26, 27, 29, 32, 36, 37):

a-hur "palm", a-dar "horn", a-ho "mouth";

i-zen "name", i-zar "star";

u-hin "wave", u-khare "wrist";

bi-hotz "heart", be-hatz "nail", be-larri "ear".

The inflectional (case) endings of Basque also have
counterparts in the same languages (Bengtson 1993). It turns
our that the grammatical comparison is even more decisive than
the lexical, and that here too Basque is clearly closer to Dene-
Caucasian (especially Caucasian and Burushaski) than to
Afroasiatic.

Basque is an "unimportant" language, in terms of
numbers of speakers or political or economic importance. But,
for the prehistory of language, the value of Basque is
inestimable, as the sole remnant, west of the Caucasus, of a
once far more widespread family of languages: Dene-
Caucasian. The various layers of the Basque vocabulary tell a
fascinating story, which is yet to be fully told. Rene Lafon
(1949:202) distinguishes the Basque physical type ("ce type
parait se rattacher a la race paleolithique de Cro-Magnon")
from the Basque language ("apparentee aux langues
caucasiques"). "C'est une langue d'origine etrangere"
(1949:206), adopted by the ancestors of the Basques, just as
their counterparts in western Iberia and Gaul adopted Celtic,
and later Latin, along with corresponding cultural and social

MOTHER TONGUE

Issue 22, May 1994

changes.
NOTES

1) For most linguists in the first half of this century,
"Caucasian" or "Caucasic" included Kartvelian ("South
Caucasian") as well as Northwest Caucasian and Northeast
Caucasian. In recent years, there has been a growing consensus
that Kartvelian is genetically quite distinct from (North)
Caucasian. (Linguists of the Nostratic School generally
include Kartvelian in Nostratic but include [North] Caucasian
in Dene-Caucasian.) Since the designation "South Caucasian"
has been cast aside in favor of "Kartvelian", "North" Caucasian
is now simply "Caucasian".

2) Because this is not a glottochronological
calculation, comparisons with plausible semantic shifts are
included. (The intention is to focus on basic semantic fields).
Basque-Afroasiatic comparisons are taken from Mukarovsky
(1969, 1981), and Basque-Caucasian and Basque-Burushaski
comparisons are drawn from Cirikba (1985), Blazek (1991
[=1992]), Blazek and Bengtson (1995), and Bengtson (1990;

1991a, 1991b, 1991c, 199 Id, 1991e, 1991f; 1993). Because I
consider Basque to belong to a subgroup (Macro-Caucasian) of
Dene-Caucasian, most of the comparisons here involve the
three entities: Basque, Caucasian, and Burushaski. (More
remote branches of Dene-Caucasian are cited in a few cases.)
Caucasian reconstructions and attested forms follow Nikolaev
and Starostin (1991, 1992). Unless indicated otherwise,
Basque forms are cited from more phonetically conservative
"French" dialects rather than the /z-less "Spanish" dialects:
e.g., mihi, hortz rather than mi(i) 9 ortz.

3) Basque *beryi 9 which recapitulates the Avar and

Chechen forms, is based on the Souletin form ber-phuru
"eyebrow", from *bert-buru from *bery-buru, according to
Basque rules of combination; cf. Burushaski -l-pwr "eyelash",
a compound of the same two morphemes.

4) Both Basque and Caucasian can be derived from a
form such as *welje > *wenje > *men^e > *nemje, by a series
of assimilations and dissimilations. Remote comparison with
Na-Dene (Tlingit wes ' "louse") confirms the antiquity of the
first form in the series (*we/j- or *welc'-).

5) Burushaski t- 9 -It- are the regular correspondences
to Caucasian A-, -K-. The Basque form was probably originally
*a-rdar or *a-ldar 9 simplified to a-dar, perhaps influenced by
Egyptian tar (§28a).

6) The higher score of Caucasian as opposed to
Burushaski can be explained by two factors: (a) the large
number of Caucasian languages, and (b) the relatively scant
amount of Burushaski material available to linguists.

7) That is, Basque parallels with Kartvelian. The
direction of loans is not always clear. Most probably, they
result from a period when the linguistic ancestors of the
Basques lived in geographic proximity to the Kartvelians (in
the area now occupied by the Ossetes?).

8) French or, Spanish oro "gold", etc. (from *auso-)
are true "chance resemblances"!

REFERENCES

Bengtson, John D. 1990. "An End to Splendid Isolation: The

Macro-Caucasian Phylum". Mother Tongue 10 (April

1990).
Bengtson, John D. 1991a. "Notes on Sino- Caucasian", in

Shevoroshkin (ed.) 1991, pp. 67-129.
Bengtson, John D. 1991b. "Some Macro-Caucasian

Etymologies", in Shevoroshkin (ed.) 1991, pp. 130-141.
Bengtson, John D. 1991c. "Macro-Caucasian Phonology (Part

I)", in Shevoroshkin (ed.) 1991, pp. 142-149.
Bengtson, John D. 199 Id. "Postscript I. (Macro-Caucasian

Again)", in Shevoroshkin (ed.) 1991, pp. 150-156.
Bengtson, John D. 1991e. "Macro-Caucasian: A Historical

Linguistic Hypothesis", in Shevoroshkin (ed.) 1991, pp.

162-170.
Bengtson, John D. 199 If. "Macro-Caucasian Again". Mother

Tongue 13:19-26. (April 1991.)
Bengtson, John D. 1993. "The Macro-Caucasic Hypothesis".

Dhumbadji! 1/2:3-6.
Blazek, Vaclav. 1991. "Basque and North Caucasian or

Afroasiatic?" Mother Tongue 14 (August 1991).
Blazek, Vaclav. 1992. "Basque and North Caucasian or

Afroasiatic", in Komparative Afrikanistik (Sprach-,
gesichts- und literaturwissenschaftliche Aufsdtze zu
Ehren von Hans G. Mukarovsky). Vienna.
Blazek, Vaclav and John D. Bengtson. 1995. "Lexica Dene-

Caucasica". Central Asiatic Journal (forthcoming).
Cirikba, Vjaceslav A. 1995. "Baskij i severokavkazskie

jazyki", in Drevnjaja Anatolija, pp. 95-105. Moscow:

Nauka.
Dolgopolsky, Aaron B. 1964. "Gipoteza drevnejsego rodstva

jaykovyx semej Severnoj Evrazii s verojatnostnoj tocki

zrenija [A Probabilistic Hypothesis Concerning the Oldest

Relationships among the Language Families of Northern

Eurasia]". Voprosy Jazykoznanija 2:53-63. [English

translation in Shevoroshkin and Markey (eds.) 1986:27-

50.]
Lafon, Rene. 1949. "Sur les origines des Basques et de leur

langue". Les Cahiers d'Outre-Mer 7:193-207.
McCall, Daniel and Harold Fleming. 1994. "The Pre-Classical

Circum-Mediterranean World: Who Spoke Which

Languages?" Mother Tongue 21:22-29.
Mukarovsky, Hans G. 1969. "Baskisch-berberische

Entsprechungen". Wiener Zeitschrift fur die Kunde des

Morgenlandes 62:32-51.
Mukarovsky, Hans G. 1981. "Einige hamitosemitische und

baskische Wortstamme". Berliner afrikanistische

Vortrdge (Serie A: Afrika, Band 28), pp. 103-1 18.
Nikolaev, Sergei L. and Sergei A. Starostin. 1991. "North

Caucasian Roots", in Shevoroshkin (ed.) 1991, pp. 174-

264.
Nikolaev, Sergei L. and Sergei A. Starostin. 1992. A

Caucasian Etymological Dictionary. Manuscript.
Shevoroshkin, Vitaly, ed. 1991. Dene-Sino-Caucasian

Languages. Bochum: Brockmeyer.
Shevoroshkin, Vitaly and Thomas L. Markey, eds. 1986.

Typology, Relationship, and Time. Ann Arbor, MI:

MOTHER TONGUE

Issue 22, May 1994

Karoma.
Starostin, Sergei A. 1989. "Nostratic and Sino-Caucasian", in
Vitaly Shevoroshkin (ed.), Explorations in Language
Macrofamilies, pp. 42-66. Bochum: Brockmeyer.

LETTER FROM JERRY KING TO HAL FLEMING

O siyu
Greetings:

ZJC

Hal, thanks for everything and your continued interest
— a letter to follow when I'm feeling better. Quickly did these
two notes. Thought you might find this interesting, as it
represents salvage work. Did this while recovering from
surgery. Had to use what was near. Pardon my shaky hand and
scratch paper, but wanted to get ideas down. Don't know if this
is what you want for Mother Tongue or not, but thought you
might like to see the notes as it fills a gap in Americas. More
when feeling better.

-C# Tohi

peace and health".

Sincerely, Qff^^J

A NOTE ON OFO s/ca/o "HEAD"

JERRY KING
Lavonia, GA (U.S.A.)

I had been studying the Ofo language, a Siouan
language of the Southeastern U.S.A. for a long time, when I
happened to meet the late Claude Medford, Jr., a Choctaw, in
1988. Claude had had the good fortune to know very well
Alice Picote (Tatoo, her name in Ofo), the granddaughter of
Rosa Pierrette (Little Flint, her name in Ofo), the informant of
J. R. Swanton, who saved the Ofo language for the scientific
world. Claude became intimate friends of Alice and made a pot
in the shape of a human head with tatoo marks at Alice's
request to be buried with her on her death. He and Chief
Joseph of the Tunica buried her, honoring her request. Claude
gathered much additional information on the little documented
southeastern tribe, and gladly we exchanged our information,
which led to some important discoveries.

Greenberg (1987:172) correctly shows two competing
forms for "head" in his subgroup: Almosan-Keresiouan of
Northern Amerind. Ofo seems to have preserved cognates of
both forms. Most of the Siouan languages have cognates of the
Proto- Siouan word for "head" from *pa, indeed the most
common words for "head" in Ofo are descended from this
Proto-Siouan word. They are apha, -pha, and -pa. Greenberg
does not list these Ofo cognates. However, Ofo seems to have
preserved a cognate from the second proposed term for "head"
which in Ofo is skdlo (Dorsey and Swanton 1912:329). It

seems to share this cognate with a closely related language
Tutelo, which, together with Biloxi, make up what has been
termed Ohio Valley Siouan by Voegelin (1941:246-249) and
Southeastern Siouan by Haas (1968:84). The Tutelo cognate
listed by Greenberg is sako = "above". Also, another Siouan
language, Catawba, appears to have a cognate which is listed
by Greenberg as -ska = "head". All of these in turn appear to
be related to Cherokee, listed by Greenberg u-sko-li, -sk- =
"head". For those interested in deeper relation, Greenberg
considers Wakasan Kwakiutl saq ? a = "over" and Nisqualli, a

Mosan Salish language, suk, sisuk = "above" as cognates as
well.

Claude remarked that Alice referred to the U. S.
currency quarter as "double head" because of the eagle head on
one side and Washington's head on the other. Fortunately,
Swanton gathered from Alice's grandmother Rosa the term for
quarter as skdlo ndpha = "two head". Whether Rosa fully
explained this to Swanton or not I do not know as I have not
seen his field notes. At any rate, Swanton gave the etymology
of skdlo as a translation of French escalin, the French coin
valued at 1214 cents or, in English, a "bit". He implies that it is
a borrowing from French. But note that he does not say it is a
borrowing. Nor have I found where he did. Both Claude and
I rejected the idea that it was a borrowing and firmly believed
it to be a true cognate with Cherokee and Catawba. We both
believed it to be a "bit" (pardon the pun) of folk etymology
here.

There are several other reasons for rejecting it as a
borrowing. Alice also referred to the U. S. currency gold coin,
commonly referred to in English as the "double eagle" as
"double head". This certainly suggests a much wider usage
than the French escalin or 1214 cents.

Further, the name "double head" was and is a popular
personal name among southeastern Native Americans. This
could easily account for its persistence in Ofo as an archaic
term. "Double head" or tal tsuska, tar tsuska is a very common
Cherokee personal name, and the name of a very famous
Cherokee Chief. The name has deep metaphysical meanings.
It is also a common art motif, deeply rooted in Native American
art from early Adena times (ca. 1,000 BC) and attested in art
forms since then by numerous artifacts. One further note here,
one still sees in print that this design was borrowed from the
European royal coats-of-arms, especially the Hapsburgs. This
is simply not the case, as it is attested for over 3,000 years in
eastern North America and perhaps the northwest U.S.A. as
well.

Since there is a scarcity of Ofo material and the
difficulty of using what has been preserved, the discovery of a
cognate skdlo = "head" in Ofo to other Almosan-Keresiouan
terms should aid us in better understanding the evolution of the
speech of Native America as well as the evolution of the speech
of the human family.

Northern Amerind

A. Almosan-Keresiouan
1 . Almosan
(a) Algic

MOTHER TONGUE

Issue 22, May 1994

(b) Kutenai

(1) Chemakuan

(2) Salish:

Nisqualli: suk, sisuk "above"
(3) Wakasan:

Kwakiutl: S9q ? a "over"

2. Keresiouan

(a) Caddoan

(b) Iroquoian:

Cherokee: u-sko-li, -sk- "head"

askoli, uska, -sk- "head"

(d) Siouan- Yuchi

(1) Yuchi

(2) Siouan:

Catawba: -ska "head"; sako, sag- "above"

iska "head"
Ofo: skdlo "head" (archaic form competing

with -pha-, -pa- [the more common

form])
Tutelo: sako "above"

B. Penutian

C. Hokan

REFERENCES

Dorsey, J. O. and J. R. Swanton. 1912. A Dictionary of the
Biloxi and Ofo Languages. (= Smithsonian Institute,
Bureau of American Ethnology, Bulletin 47.)
Washington, DC: Smithsonian Institute.

Greenberg, Joseph H. 1987. Language in the Americas.
Stanford, CA: Stanford University Press.

Haas Mary R. 1968. "Last Words of Biloxi". International
Journal of American Linguistics. 34/2:77-84.

Voegelin, C. F. 1941. "Internal Relationship of Siouan
Languages". American Anthropologist 43:246-249.

A NOTE ON CATAWBA Weyaline
"CHIEF'S TOWN"

JERRY KING
Lavonia, GA (U.S.A.)

It is well known in scientific circles and the general
public as well that the greatest linguistic diversity in aboriginal
North America was found in the California-Oregon area, as far
as we presently know. Dr. Mary Haas made this point in 1970
(1971:44) and pointed out the reason was that his area,
especially the northern part, was largely little affected by the
inroads of European civilization until the middle of the
nineteenth century.

At that same time, Dr. Haas also pointed out to
scientific circles that there was another area of great linguistic
diversity in the native Americans and that it was certainly in the

southeast U.S.A. and the adjacent coast of the Gulf of Mexico.
A fact well known among knowledgeable Native Americans.
She pointed out that we may never know just how diverse this
area really was because of the sparseness of information,
especially so for the critical period before Native Americans
had been seriously dislocated by the pressures of competing
European nations. It is conceivable that it just may have been
more diverse, possibly the most diverse in aboriginal North
America. Because of the lack of data, however, we may never
know. As she pointed out, "many smaller tribes in the
Southeast [U.S.A.] have almost certainly vanished without a
trace. Others are known to us by names only..." (Haas
1971:44). And I should like to point out that, even among well
documented languages of the Southeastern United States, there
are serious gaps in documented dialectal variations and archaic
forms, as many of which have not survived into modern times.
These gaps are sometimes a serious handicap in reconstructing
accurate proto-words and, therefore, seriously handicap
comparative work. These problems are not only of concern to
Americanists but others, especially "long rangers". So it
behooves us to use every means possible to aid in overcoming
these serious problems. It is to these concerns that I address
these notes.

Recently, I was shown a photocopy of a map dated
1756 of the Cuttahbaw Nation (the modern Catawba Nation of
South Carolina, U.S.A.). They have recently been in the news
media as they have settled a land claim and reestablished a
federal relationship (reservation, etc.). The map shows several
Catawba towns, and a number of place names can be recovered
from it. One of the towns is labeled "Weyane or the King's
Town". The name of the town is repeated two more times on
the map, labeling "trails" to the town itself and both times
spelled "Weyanne". I have found the name of the town again
in John Evans Journal to the Catawba Nation. Here it is as
"Weyaline". Again with these words "the Catawbaw King... his
town" (McDowell 1970:107). We thus have three attested
forms, Weyaline from the text and Weyane and Weyanne from
the map. And we are told it translates as "Chiefs Town" or
"King's Town" (i.e., "Kingston") in English. Immediately, one
knowledgeable of Catawba will recognize that the first syllable
we does indeed mean "town". This then leaves us with the
syllables -yaline, -yanne, -yane, which we are told mean
"chief or "king". Here we have a problem as in modern
Catawba the word for "chief is ye swf and migmehe which
does not match, at least in modern Catawba (Speck and
Schaeffer 1942:564). Thus, it would appear "Chiefs Town" is
not accurate, or is it? Here, comparison is of value. Catawba
is a Siouan language, eastern division. In Ofo, another Siouan
language, western division, the word for "chief is itcole (<
iyole in Archaic Ofo). Here, we have a possible match,
especially with the archaic form. If Ofo iyole and Catawba
-yaline are cognates, then Weyaline does probably mean
"Kingston", and -yaline is an archaic form in Catawba.

Is there any supporting evidence to strengthen this
argument? Again, we turn to old documents. The South
Carolina Gazette, published on June 2 and on June 7, 1746, call
the Catawba King Yanabe Yalengway. Note the syllables in the
second word Yalen- and note also he is referred to as "The

MOTHER.TONGUE

Issue 22, May 1994

Catawba King" (Brown 1966:224). Hyaline in the Evans text
means "king" in 1756 and variations yane andyanne appear in
the same year of 1756 on the map with the meaning "king", and
we now have ten years earlier in 1746 in the South Carolina
newspapers yalen, a part of the Catawba King's Name and a
likely cognate in Old Ofo iyole meaning "chief, then it is very
probable that yaline is an old Archaic Catawba word meaning
"king, chief, leader" based on a root other than the modern
forms recorded.

One other note: I collected the Catawba term lyurenal
which my Catawba informant glossed as "my lead" (i.e.,
"leader") in reference to a black and tan hound which was the
"lead dog" of his hunting pack of "coon dogs". He called his
favorite "lead dog" by the Catawba name /wadesj/ or "old
gourd" from wade = "gourd" and sj = "old" because of the
dog's long nose and the black and tan color. Now compare his
reference to the dog as yurena = "my lead (dog)" based onyure
= "lead" and -na = "my". This suggests a very good translation
of weyaline as we = "town", yali = "leader", ne = "my". If
Evans had a translator or interpreter, which he stated he did,
then this translator would have probably responded to Evans
inquiry with the Catawba term weyaline or "my leader's (i.e.,
chief or king) town". This certainly compares nicely with
yurena, "my leader".

My present thinking is that we have recovered a
Catawba term derived from an Archaic Catawba term which
fell into disuse as a political term in modern terms and offers us
an opportunity to project back 4,000 to 6,000 years ago to
Proto-Siouan times and a new piece to the puzzle of what
Proto-Siouan contained with the reconstructed Proto-Siouan
term *yale = "to lead; leader".

This, then, permits us to look at an even deeper
connection of 6,000 to 8,000 years ago to Macro-Proto-Siouan
times. For the Proto-Siouan term *yale = "to lead; leader" may
ultimately be connected at least to the *ya syllable, to my own
Cherokee suffix -ya found in the tribal name of the Cherokee
Yvwiya from Yvwi = "human being" and -ya = "principal,
original, etc." We now have a Macro-Proto-Siouan term *ya
meaning "lead, principal, original, etc."

In these notes, I have used numerous extra-linguistic
materials , such as old maps, journals, newspapers, place
names, field notes of survivors, etc., as well as linguistic data.
In this search, no possible lead must be overlooked. Research
of this nature must, to be fruitful. I grant that much work needs
to be done: the exact phonetics involved, more examples,
especially of archaic forms and dialectal variations, but we
must have the data to examine, if we are to proceed in unveiling
the past and ultimate origins of our speech.

REFERENCES

Brown, Douglas Summers. 1966. The Catawba Indians. The
People of the River. Columbia, SC: University of South
Carolina Press.

Haas, Mary R. 1971. "Southeastern Indian Linguistics", in
Red, White, and Black: Symposium on Indians in the Old
South. (= Southern Anthropological Society Proceedings,
no. 5.) Athens, GA: University of Georgia Press.

McDowell, W. L., ed. 1970. Colonial Records of South
Carolina Relating to Indian Affairs. Columbia, SC.

Speck, F. G. and C. E. Schaeffer. 1942. "Catawba Kinship and
Social Organization with a Resume of Tutelo Kinship
Terms". American Anthropologist (part 1, Oct.-Dec.)
N.S. 44:4.

ALTAIC, GERMANO-EUROPEAN,

AND NOSTRATIC:

THE EVIDENCE OF PHONETICS AND

PHONOLOGICAL SYSTEMS

TOBY D. GRIFFEN

Southern Illinois University at Edwardsville

1. Phonetics, Phonological Systems, and Nostratic

The bulk of the argumentation for or against the
Nostratic theory has centered on shared lexical items which
appear to have corresponding relationships among the sounds,
particularly among the occlusive obstruents. Thus, for example,
Kaiser and Shevoroshkin (1988 - following Dolgopolsky
1964) note the correspondence between Indo-European *k w er-
'to build' and Altaic *Kurv- 'build', in which both the
obstruents and the trills correspond. In other instances,
correspondences appear to have been affected by intervening
sound laws, as in Indo-European *g w en- 'woman, wife' and
Altaic (Turkic) *kuni 'one of the wives (in polygamy)'.

One aspect of the correspondences that has received
little attention, however, is how the corresponding obstruents
operate within their systems. It is not enough to establish
correspondences between obstruents in isolation, for the
"same" obstruents may or may not correspond, depending upon
the systems in which they occur.

1.1. Borrowing between Systems

When languages (or more properly, dialects) come
into contact with one another, there is usually some degree of
lexical borrowing. When a word in one language is borrowed
into the other, though it is pronounced with the "accent" of the
borrowing language and it is thus incorporated into the
borrowing lexicon with the phonetic characteristics of the
borrowing language and within that language's own
phonological system (compare Fries and Pike 1949).

Occasionally, a sound is borrowed from one language
into the other in order to accommodate the lexical item being
loaned. For example, it has been demonstrated in Griffen 1974
that the affricates were borrowed from English into Welsh to
render certain English words. Once they were incorporated
within Welsh, however, they ceased to be treated as English
sounds and adhered to the phonological system of Welsh.

The development of the Welsh affricates has taken
place on two levels. On the phonological level, the tenuis

MOTHER TONGUE

Issue 22, May 1994

affricate [c] now alternates with the media affricate [jf] in the

mutation system just as the tenuis stop [t] alternates with the
media [d]. Moreover, a series of nasal affricates [nh/ii] has

developed to accommodate nasal mutation in the dialect of
Dyffryn Nantlle (see also Jones 1971). Thus, the incorporation
of the affricate into the Welsh phonological system is complete
only when it functions within the morphonological alternation
patterns of the language.

It is on the phonetic level, however, that the initial
incorporation must take place. While the consonants of English
are differentiated by voice, those of Welsh adhere to a four-
member aspirate fortis-lenis scale. It would be quite impossible
for the affricates as English sounds within the English system
to participate in the mutation system based upon the fortis-lenis
scale. In borrowing the tactic of affrication, Welsh thus alters
not its system, but rather the perception of affrication from that
of a voicing system to that of an aspirate fortis-lenis system.

In borrowing, the phonetic underpinnings of a
phonological system do not change just because a foreign word
is incorporated into the lexicon. Indeed, as demonstrated in the
phonemic categorization experiments (see especially Scholes
1968), no matter how well the speaker knows the foreign
language, the speaker perceives the sounds only within the
system of the native language.

For example, speakers of Swabian, an Alemannic
dialect in south-western Germany, have long had contact with
French, at least through the nearby Alemannic dialect of
Alsatian. Nonetheless, the French word pardon [pardon] has
been borrowed as [bardo] by the Swabians (see Kauffmann
1980:176, Griffen 1983:151-52). The tense voiceless
unaspirate labial stop within the French obstruent system based
upon muscle tension is interpreted as the lax voiced unaspirated
labial stop within the Swabian obstruent system based upon
fortis aspiration. While the Swabian [p h ] would have actually
been closer to the French [p] in its phonetic features, the
number of corresponding features is not what is important -
what is important is the correspondence of (un)aspiration alone,
for that is the pertinent opposition in the borrowing system.

1.2 Sound Correspondences in Nostratic

Within the comparative method, Nostraticists have
relied upon sound correspondences between words in the
various language groups, as pointed out in section 1. Thus,
Indo-European *k w er- 'to build' is seen to be related with
Altaic *Kurv- 'build', partly because the dorsals in both
correspond with one another, and the lexical items are close
enough in meaning to have been derived from a common
source. Moreover, such researchers as Kaiser and
Shevoroshkin (1988) as well as Bomhard (1992 - see also
Bomhard and Kerns ms.) demonstrate that the corresponding
sounds can be related by sound laws, just as are the
corresponding sounds among the various Indo-European (or
Germano-European) language groups.

Critics of the Nostratic theory point out that such
correspondences and the sound laws governing them may well
be the result of borrowing. For example, the fact that the
consonants in German Strafie [Jtra:s9] 'street' can be shown to

correspond through regular sound laws with those of Italian
strada [stra:da] 'street' does not demonstrate a common
ancestor at all. German simply borrowed the word from Latin
at a time prior to a relatively late sound shift.

Certainly, it is difficult to determine whether sound
correspondences among languages at the (pre-)historical
"depth" of the Nostratic language families may result from
borrowing or from genetic relationship. Thus, Campbell
(1990:174-75) concludes his study of similarities between tree
names in Indo-European and Finno-Ugric (or Uralic) as
follows:

In summary, the two language families, FU (or U) and
IE share a rather large number of similarities among
their names for trees. While conceivably some of these
compared forms are but fortuitously similar, the
weight of the aggregate of comparisons is sufficient to
support the conclusion that these two language
families have a very old historical connection, one
which reflects either a genetic affiliation or
Sprachbund affinities, or perhaps both. The
connection may involve diffusion, and indeed, certain
of the forms presented here almost certainly involve
borrowing. Given this set of circumstances, the
possible areal linguistic relationship to explain these
and other observed similarities among some of the tree
names explored here may perhaps reflect an old
genetic relationship - a common ancestor. Further
study ought to keep both hypotheses open, and it is
hoped that the comparisons presented here will indeed
stimulate further research aimed at determining the
exact nature of the historical connection shared by IE
andFU(orU).

Thus, it is just as easy for proponents of the Nostratic
theory to see sound correspondences among the pertinent
families as evidence of genetic relationship as it is for
opponents of the theory to see the same data as evidence of
borrowing. A greatly neglected fact about borrowing though
(see above and Griffen 1974), is that the sounds would have to
have been incorporated within the borrowing phonological
system based upon the phonetic characteristics of the
borrowing language. In no event could the phonetic basis and
its phonological system have been borrowed.

The question now for historical linguists, whether they
support or oppose the Nostratic theory, is this: Are the
phonological systems of the Nostratic cluster of language
families and their phonetic bases close enough to one another
to suggest a genetic relationship? Given the many possibilities
for sound systems available to unrelated language families, the
occurrence of regular sound correspondences between similar
phonetic/phonological systems would make a strong case for
the genetic relationship. Conversely, the occurrence of regular
sound correspondences between significantly different
phonetic/phonological systems would make a strong case for
borrowing.

This article will provide a starting point for the answer
to this question. With the new reconstruction of Germano-
European (Griffen 1988, 1993) rather than the traditional Indo-

MOTHER TONGUE

Issue 22, May 1994

European, the relationship between this family and Nostratic
will be examined (as in Griffen 1989) to see how closely the
phonological systems and their apparent phonetic bases
correspond. Next, the obstruent system of Altaic will be
examined to see if there is indeed an argument for systematic
correspondence between at least these two hypothesized
Nostratic families.

2. Germano-European and Nostratic

German-European (See Griffen 1988) is a term
applied to a reconstruction of the proto-language consonant
system for the language groups heretofore known as Indo-
European. While the Indo-European reconstructions have been
carried out in the old segmental phonetic/phonemic tradition,
the Germano-European reconstruction has been accomplished
through the more modern dynamic analysis (see Griffen 1985).
Based upon a more reliable system of analysis without the
prejudices of segmental letters, this reconstruction maintains
that the development of the proto-language obstruent system
was rather simple in Germanic, but indicative of a more
complicated shift in the "Indo-European" (non-Germanic)
language groups.

2. 1 Phonetic Plausibility

Moreover, the reconstruction inhances its reliability
through the theory of phonetic plausibility (Griffen 1993). An
outgrowth of Prague School phonology with its insights into

the way feature/oppositions are related through the
phonologically pertinent opposition (see especially Trubetzkoy
1969:76-77, Jakobson 1962:23), this theory maintains that
predictions and reconstructions of change must adhere to the
functioning of the system and to the pattern of change inherent
to it. For example, in a lenitive environment (word-final,
intervocalic, etc.) in an aspirate fortis-lenis system with [t] as
fortis and [d] as lenis, a lenitive change (from [t] to [d]) is
plausible, while a provective change (from [d] to [t]) is
implausible. Since the consonants among the language groups
correspond along the fortis-lenis scale, the provective Grimm's
Law is implausible, while the Germano-European
reconstruction is plausible, for it adheres in all instances to
known fortis-lenis relationships.

The Germano-European reconstruction is based upon
the phonetically plausible developments in an aspirate fortis-
lenis system. In such a system, a fortis-lenis scale such as that
characterized by [6] - [d] - [t] - [9] from lenis to fortis is based
upon the phonetic parameter of fortis aspiration, characterized
by an increasing ratio in high-to-low frequency emissions from
lenis to fortis (see Griffen 1975a:chapter 10, 1985:chapters 5
and 7, 1993). As seen in historically attested changes in Welsh,
this type of system can support the following expectations for
change (all quite plausible, but listed in descending order of
plausibility):

1. Where accent is not a factor, intervocalic consonants may
change to the lenis.

2. Where accent creates a position of fortis strength,

Table 1: Germano-European Reconstruction

Reconstruction

* P -

p h et- 'foot'

*-p h -

nep h od- 'grandchild, nephew'

*P"

paita- 'goat skin'

*- P -

deu-p- 'deep'

*- P -

uper 'over'

*b-

ber- 'to carry'

*-b-

terb- 'to wind'

* t h _

t h rei- 'three'

*_t h _

uet - 'year'

*t-

tuo(u) 'two'

*-t-

set- 'to sit'

*-t-

k mtom 'hundred'

*d-

de(i)- 'to nurse'

*-d-

beud- 'to observe'

*k h -

k h uon- 'dog'

*-k h -

t ak - 'to hush'

*k-

kar- 'to call, to cry out'

*-k-

sak- 'to track'

*-k-

sek- 'to cut'

*g-

ger- 'to jut forth'

*-g-

steig- "to climb'

Germanic

Gotfotus'foot'(3)
OIc nefe 'nephew' (3)
Got paida 'garment'
Gotdiups 'deep'
OHGubir'over'(l)
Got bairan 'to carry'
OHG zerben 'to turn'

Got J^reis 'three' (3)

Got wi^rus 'yearling' (3)

Got twai 'two'

Got sitan 'to sit'

Got hund 'hundred' (1)

Got daddjan 'to nurse'

OS biodan 'to present'
Gothunds'dog'(3)
Got {)ahan 'to hush' (3)
Gotkara 'care'
Got sokjan 'to seek'
Got sega 'saw' (1)
OIc grgn 'mustache'
Got steigan 'to climb'

Indo-European

Lat pedis 'foot (gen.)'

OP napat- 'grandchild'

Grk (3aixr| 'goat(skin) garment' (1G)

111 bythos 'deep' (1)

Av upara 'the upper' (2/1)

Sktbharati 'carries' (1G)

Skt drbhati 'winds' (1)

Olr tri 'three'

Hit witt- 'year'

Albdu'two'(lG)

OCS sedeti 'to sit' (1)

Lat centum 'hundred' (2/1)

Skt dhaya-h 'nourishing' (1G)

Skt bodhar 'knower' (1)

Toe ku 'dog'

Umb ta^ez 'silent'

Grk yfipix; 'voice' (1G)

Olr saigim 'to seek' (1)

OCS sekyra 'ax' (2/1)

Grk %apf|v 'upper lance point' (1G)

Skt stighnoti 'climbs' (1)

MGTHEK tongue

Issue 22, May 1994

intervocalic consonants either do not change, or change to
the fortis.
3. A consonant in a position of strength (as in word-initial
position) may change to the fortis. (Griffen 1988:174,
also 1989:143)

From this phonological system with its aspirate
phonetic basis, the reconstruction of Germano-European can be
accomplished as in table 1 (p. 40), in which the numbers refer
to the tendencies noted above and (G) to a generalization of a
tendency to a new environment. All reconstructions are thus
phonetically plausible in keeping with the phonetic basis of the
phonological system.

The original Germano-European obstruents thus
maintained three members: (1) a strongly articulated aspirata
*r, (2) a less strongly articulated (less aspirate) tenuis *t, and
(3) a more weaky articulated (relatively least aspirate) media
*d. From this reconstruction, the developments in all the
families are well within expectations, as an aspirata may be
provected to a spirant (in Germanic initials) or lenited to a
tenuis (in Indo-European), a tenuis may be lenited to a media
(in all - the strongest tendency), and a media may be lenited to
a susurrata (in Indo-European and later in Germanic as well -
the term susurrata is used for a glide, murmur, voiced fricative,
or other such continuous "whispered" weakly articulated
obstruent in a series - see Griffen 1988:24).

2.2 Correspondences with Nostratic

This reconstruction of Germano-European consonants
was conducted without any particular knowledge of the
Nostrastic theory. Nonetheless, the Germano-European
reconstruction fits far more precisely into the Nostratic
reconstruction of Kaiser and Shevoroshkin (1988) and of
Bomhard (1992 - also Bomhard and Kerns ms.) than do any of
the Indo-European reconstructions. The Nostratic, Germano-
European, and Indo-European reconstructions are summarized
in table 2.

Table 2: Occlusive Obstruent Reconstructions

Nostratic

Germano-
European

Indo-European

**p'
**p

**b

*p h

*b
*bh

**t

*t h

*t
*d
*dh

*k h

*g
*gh

As seen in table 2, the Germano-European and
Nostratic reconstructions are practically identical in their

occlusive obstruent series. The major difference is in the fortis-
most obstruent, realized in Germano-European as an aspirata
and in Nostratic as a glottalic. Since the glottalic feature
involves a forced glottal ejection and the aspirate feature
involves an emphatic glottal frication, that the former should
give rise to the latter (or even vice versa) is not at all
implausible.

The only other possible point of contention with the
adherents of the Nostratic theory may be in the position of the
media in Germano-European. Kaiser and Shevoroshkin
(1988:323) consider the lenis-most realization of the obstruent
in this family to be "breathy-voiced," to support borrowings
from Semitic into Indo-European and from Indo-European into
Kartvelian. Since the parts of the family in contact with these
other families would have been the non-Germanic, postshift
Indo-European groups though, this observation of borrowing
patterns does not detract from the Germano-European
hypothesis; nor does the Germano-European hypothesis detract
from the Nostratic.

The obstruent system of Germano-European thus
corresponds as precisely as could be expected with that of the
Nostratic reconstruction. The phonetic basis in glottalic activity
for the Nostratic reconstruction is quite compatible with the
fortis aspiration of the Germano-European. Indeed, the same
tactics may be expected to apply in both, at least to a
significantly overlapping degree. Systematically, if the
Nostratic theory were proven, then the hypothesis that
Germano-European could be a member of that macrofamily
would be very strong indeed - supported both by regular sound
correspondences, which could come from borrowing or genetic
relationship, and by the shared system, which could not be
borrowed, but would be most likely a result of genetic
relationship. Of course, there is the argument for a coincidence,
but given the many possibilities for a phonetic basis of an
obstruent system, this would be improbable; and it would be
made even less probable by the extent of the contact areas
between Germano-European and Nostratic.

3. The Altaic Obstruents

The primary source for this examination of the Altaic
obstruent system is Poppe (1960:9-25, 42-62), drawn from
many sources, the most important of which include Ramstedt
1957 and Vladimircov 1929 (compare also Shirokogoroff
1931). While there is considerable debate on the members of
the family, the families that are consistently included in Altaic
are the Mongolian, Manchu-Tungusan, and Turkic. Korean is
increasingly included in Altaic (as Poppe does), and there has
been some argument for Japanese (compare Miller 1971).
Following Poppe (1960:8), the relationships among the Atlaic
language groups can be visualized as in table 3 (p. 42).

The fortis-most obstruents of proto-Altaic are now
consistently reconstructed not as tenues, but as aspiratae.
Among Nostratic researchers, Kaiser and Shevoroshkin (1988)
represent the fortes as markedly aspirate, and Poppe notes that
"On the occlusives and affricates, it should be noted that they
were differentiated not so much between voiceless and voiced
as between strong (fortis) and weak (lenis). The strong

MOTHER TONGUE

Issue 22, May 1994

Table 3: Altaic Language Relationships

"Greater"

Altaic

proto-Korean

pre-Turkic

pre-Mongolian/Manchu/Tungusan

proto-Turkic

proto-Chuvashic
Chuvashic

proto-Mongolian
Mongolian

proto-Manchu-Tungusan

Turkic

Manchu

Tungusan

Korean

occlusives (*/?, *t, *q, *k) and the strong affricate (*c) were

probably strongly aspirated" (1960: - my translation from the
German). While Poppe usually represents the fortis labial
obstruent, for example, simply as */?, the practice in this
examination will be to represent it as p (or [p ]), adjusting
Poppe 's transcriptions accordingly.

The following examination of the Altaic obstruents is
conducted through the three degrees along the fortis-lenis scale

- aspirata, tenuis, and media. All examples are taken from
Poppe (1960) and are cited simply by page number in
parentheses.

3. 1 The Altaic Aspirata

Proto-Altaic maintained three plosive fortis obstruents

- *p h , */*, and *A^. These all occurred in initial position and
were evidently heavily aspirated, as seen in the changes they
underwent in the development of the various Altaic languages.
The affricate *c was probably also fortis and may well have

patterned itself after the other obstruents, but its development
does not shed any light upon the nature of the Altaic obstruent
system under study and is therefore not considered here in any
detail.

The reflexes of the three Altaic plosive obstruents are
listed in table 4. All of these occurrences are in initial position,
which tends to be marked in an aspirate system for heavy
aspiration (see section 2.1). The fact that the aspirata is found
only here, of course, suggests what has traditionally been seen
as allophonic variation and what is more precisely the
coarticulation of an obstruction with an aspirate environment
(compare Griffen 1985:chapter 5). This suggestion is addressed
in section 4.

3.1.1 Labial. By far the fortis obstruent that is most
supportive of the relationship with Germano-European is the
labial. Through this, it is also most clear why the fortes are
considered to have been heavily aspirated. As seen in table 4,
the aspirated fortis *p is realized in several ways in the Altaic
languages.

In Manchu as well as in Monguor (Mongolian), the
change from *p h to /is typical of provection (or fortition - a
change toward the fortis) in an aspirate fortis-lenis system and
parallels precisely the development in Germanic. As noted in
Griffen 1988, such a change would necessarily have come from
an aspirate obstruent in a fortis (initial) position. Thus, the
correspondence between Goldic (Nanaj) para 'sled' and
Manchu far a 'sled' (11) parallels that between Latin pater

Table 4: Altaic Aspirata

Labial - *p

Dental - *t h

Dorsal - *k

proto-Mongolian

*p h

proto-Mongolian

*t h

proto-Mongolian

*q h3 *k h4

Middle Mongolian

Written Mongolian

q\k 4

Monguor

f 1

Manchu

Tungusan

k,g 5

East Mongolian

null

Tungusan

pre-Turkic

*q\ *k 4

Manchu

Turkic

Chuvashic

x\k 4

Goldic

SW Turkic

d 2

Old Turkish

3 i 4

q,k

Evenkic

Korean

t&th

Korean

pre-Turkic

proto-Turkic

SE & SW Turkic

elsewhere

null

Korean

p &ph

l)in other positions x, etc. 2) and sporadically in other dialects 3) postvocalic 4)prevocalic 5) before r,/

MOTHER TONGUE

Issue 22, May 1994

'father' and Gothic fadar 'father'.

One could argue that the change from [p ] to [f] could
be a lessening of occlusion and therefore a case of lenition, as
it is in the Irish mutation system, as opposed to a case of
provection as it is in the Welsh mutation system. In Irish,
however, the mutation only occurs in historically intervocalic,
lenis environments - not in all initial positions. What is at work
in Altaic is not a system based upon degree of occlusion, but
one based upon aspiration.

Indeed, the realization of h in Middle Mongolian
could also come only from an aspirate [p ] undergoing
provection, for example in the development of proto-Altaic
*p okar 'short' into Middle Mongolian hoqar (11). This
development also has its Germano-European parallel in
Armenian, in which hing 'five' corresponds with Greek nevxe.
If the starting point had been a tenuis [p], there would have
been no source for the [h] either in Middle Mongolian or in
Armenian (compare Brugmann 1972:357-58, Griffen
1988:193-94).

In the Turkic languages, the same aspiration occurred
and remains in the southeast and southwest groups. In the rest,
however, the labial obstruent has disappeared entirely, just as it
has in eastern Mongolian languages as well - for example, the
further development of Middle Mongolian hoqar 'short' as
Mongolian oqor (11). The dialects with [h] show that the same
process occurred in these as in Middle Mongolian and could
only have come from a fortis aspirata. Moreover, the ultimate
loss of occlusion has its parallel in the loss of the same
obstruent in Celtic, leading to such correspondences as Old
Irish athir 'father' and Latin pater.

As for Korean, the variation between tenuis and heavy
aspirata would certainly be consistent with the Altaic system.
Indeed, it would be consistent with an interpretation that
Korean may be the most conservative, at least in its obstruent
system (see section 4).

3.1.2 Dental. The development of the dental plosive
obstruent *r is far less interesting than is that of the labial.
Indeed, in all of the Altaic languages, the dental obstruents
resisted change the most. Of course, change is not required in
an aspirate fortis-lenis language - the patterns of change
outlined in section 2.1 are simply the way in which one expects
an obstruent system with such a phonetic basis to change, if
change occurs.

The one exception to this pattern of conservatism
occurs in the southwestern dialects of Turkic and sporadically
in others as well. This is a change from t to d - apparently a
case of lenition in a highly provective environment. Given the
strongly provective pattern of change involving the labial and
the maintenance of the fortis in the dorsal, this change is
frankly puzzling and may reflect intermediate changes not
attested in the data. While its extrasystemic behavior may mark
it as an exception (rather than as a counterexample - compare
Hjelmslev 1970:30-31) at least for the current investigation, its
occurrence in one area and sporadic occurrence in others
should warrant further study.

In Mongolian and Manchu-Tungusan, there was also
a development from the aspirate dental to the affricate before
the high front vowel, for example Mongolian cida- < *ri'da-

'to be able' (15). While one might be tempted to consider this
a case of provection, as in the High German sound shift (see
Griffen 1981b), the fact is that the same change occurred in all
positions and was doubtless due to the simple retraction of the
articulator in coarticulation with the vowel.

Once again, it is in Korean that the maintenance of the
heavy aspirate is found, alongside the tenuis (or weak
aspirata?). For example, proto-Altaic *rati 'to learn' yields
Korean that 'domesticated' (13).

3.1.3 Dorsal. While it is customary to consider this
position of obstruction dorso-velar or simply velar, it is quite
evident from the alternation between q and k that the pertinent
feature of the obstruction was its dorsality, with the precise
point of contact determined to a great extent by the vowel or by
the syllabic position (compare the German dorsal fricative in
Griffen 1977, 1985 .-chapter 3).

The single development that stands out along the
fortis-lenis scale - the realization of the plosive as a fricative
[x] - is in the Chuvashic group. Actually, this development
occurs in other Altaic languages in later stages, indicating a
consistently aspirate fortis-lenis basis to the phonological
system. For example *Icak yields Mongolian qag 'dried dirt'

but Khalkha-Mongolian xagsu 'hot, dry wind' (17); and *kf t ut-
u-k yields Mongolian qutug 'holiness' and Manchu xuturi

'holiness' (18).

This spirantization is totally in keeping with the
provective pattern in initial position of an aspirate fortis-lenis
system (the third expectation in section 2. 1). It also parallels the
development in Germanic, for example Gothic hunds <
*kruon- 'dog' (compare Latin canis, Tocharian ku).

3.2 The Altaic Tenuis

In intervocalic position and in final and
postconsonantal position, the fortis occlusives /?, t, and k were
realized as tenues. The implications for the aspirata in initial
position and the tenuis in internal and final positions is treated
in section 4.

The internal realizations of p follow two patterns -
one in a "strong" position (before a short vowel) and one in a
"weak" position (before a long vowel). These are represented
separately in table 5 (p. 44) along with the other pertinent
categories of internal and final realizations.

3.2.1 Labial. Once again, it is the labial position of
obstruction that shows the greatest degree of change and
provides the most striking evidence of the aspirate phonetic
basis of the Altaic phonological system.

Mongolian exhibits clear evidence of an aspirate
fortis-lenis system by following the first expectation of change
as presented in section 2.1 and as realized in the non-Germanic
(Indo-European) families in Germano-European. Thus, for
example, proto-Altaic *rapa yields Mongolian taba
'satisfaction' (42), with the media b replacing the tenuis/? in
intervocalic position.

Manchu and Tungusan differ in these realizations
despite their close relationship. Manchu exhibits a variety of

MOTHER TONGUE

Issue 22, May 1994

Table 5: Altaic Tenuis
Labial - *p
Intervocalic Strong

pre-Mongolian
proto-Mongolian

Written Mongolian b

Manchu f

T>
*b

Tungusan
Old Turkish

p/w
P

Intervocalic Weak

Final and Postconsonantal

pre-Mongolian

proto-Mongolian

*y, *w, *j

proto-Mongolian

Written Mongolian

Y>w,j

Written Mongolian

Manchu

-p-/-b

Goldic

p/null

Tungusan

p/w

Tungusan

p/w

Old Turkish

Korean

-w-/-p

Dental - *t

Dorsal - *k

Internal and Final

Final

Intervocalic

Written Mongolian d
Manchu t

proto-Mongolian
Written Mongolian

*q\ *k 2

proto-Mongolian
Written Mongolian

*q, *k 2
q\k 2

Old Turkish t

Manchu

q\k 2

Manchu

q ! ,k 2

Tungusan
Old Turkish

k
q\k 2

Tungusan
Old Turkish
Korean

k
q\k 2

l)posrvocalic

2) prevocalic

choices that may be indicative of a history of systematic change
(but see below). The intervocalic / could result from
gemination and its consequent provection in an aspirate system
(compare Chuvashic appa from dpd 'older sister' - 43), or it
could be indicative of a lessening of occlusion in a lenitive
environment (compare Irish, section 3.1.1). In more expected
and direct developments, the postconsonantal position does not
change, and the final position weakens to a media b, indicative
of an aspirate lenition.

The apparent confusion is cleared up by Tungusan,
however, in its variation between p and w. The maintenance of
the tenuis is quite simply no change - always a possibility in
historical linguistics. The realization of the w, however, is a
change to susurrata probably through the media b realized in
weak position (see below). This is a pattern of lenition expected
within an aspirate fortis-lenis system.

The realization off in Manchu, then, would probably
have proceeded through the w/v stage in a period of voicing.
This is precisely what happened in Latin, leaving such
correspondences as Sanskrit bhritar 'brother' and Latin frater.
(Compare Griffen 1982.)

In weak position (before a long vowel) the
intervocalic labial tenuis is realized as in table 5, above. The
variations in Mongolian are allophonic (that is, the position is
determined by the coarticulated vowel) and are not of interest
here, as the object of investigation is manner (degree along the
fortis-lenis scale) rather than position.

Here the same types of patterns emerge, with the

susurrata w (and its positional variants) realized in this highly
lenitive environment in the Mongolian languages. Hence,
pro to- Altaic *rapa 'to guess' is realized through the
intermediate *tawa as taya (47). This is consistent with an

aspirate fortis-lenis system, such as that in Germano-European.
And once again, the realization of / in Manchu is not
particularly bothersome, having a parallel in Latin, as shown
above. Nor is the ever-present possibility that no change will
occur at all.

Of particular interest, though, is the loss of occlusion
in Goldic (Nanaj). Of course, the extreme lenition through a
hypothetical media all the way to susurrata in Mongolian and
Tungusan raises the possibility that this environment is so
lenitive as to cause the ultimate lenition of the obstruent to null
(as happens in final position in the Welsh labials and dentals -
for example, the name David is realized as Dajydd [da:vid] and
ultimately as Dewi [de:wi] - note also the further weakening in
intervocalic position). In any case, the pattern of change is
lenitive in an aspirate fortis-lenis system.

3.2.2 Dental. Aside from the affrication of the dental
tenuis in Mongolian and Manchu, which is more of a positional
change and not of consequence in this investigation (see section
3.1.2), the only change in manner to be found in either the
strong position or the weak position is found in Mongolian,
where the *Ms realized in Written Mongolian as d. For
example, the name for a fire divinity *ot+kan is realized in
Mongolian as odgan (49).

MOTHER TONGUE

Issue 22, May 1994

The intervocalic lenition of t to d is well represented
in the Indo-European sound shift from Germano-European and
yields such correspondences as Gothic sitan 'to sit' and Old
Church Slavonic sedeti 'to sit' or Latin sedere 'to sit' from
Germano-European *set-. As such, it follows the most
prevalent type of change within a fortis-lenis system - the first
expectation in section 2.1.

3.2.3 Dorsal. The lenition of the internal and final
dorsal tenuis is on average about as extensive as that of the
dental, although far less extensive than that of the labial. In
final position, Altaic *k is realized in Mongolian as g or q,
variants in the postvocalic and prevocalic position within the
syllable. Thus, Altaic *cak yields (Middle) Mongolian cag

'time', Khalkha-Mongolian cag, and Sagaic sag (54). In the
other Altaic languages, the change did not occur, so that the
word for 'time' in Old Turkish is caq (54).

The realization of the media in final position is typical
of the neutralization pattern inherent to an aspirate fortis-lenis
language. For instance, the Welsh superlative form tecach
[tekax] 'fairest' corresponds with the radical form teg [te:g]
'fair'. This may be opposed to a voicing language such as New
High (East Central) German, in which the pattern is the
opposite, with the plural Tage [taiga] 'days' corresponding
with the singular Tag [ta:k] 'day'. (Compare Trubetzkoy
1969:76-77.)

As neutralization patterns in final position favor the
unmarked (or less marked) member of an opposition, the
realization of the media in this position in Mongolian clearly
indicates a lenition along the fortis-lenis scale and supports an
aspirate phonetic basis for the system. The lack of
neutralization in the other Altaic languages does not necessarily
indicate an absence of this basis - simply a nonoccurrence of
the neutralization.

In intervocalic position, however, Mongolian usually
resists the lenition of the dorsal, in spite of the fact that the
labial and even the dental have undergone lenition. There are
individual cases, however, in which the lenition does occur,
and it occurs in some rather basic vocabulary items. For
instance, the Altaic word for 'father' *etike is realized in
Mongolian as ecige (56). Such realizations are in typically

weak, lenitive environments, as in the third syllable (where
accent falls on the first). The realization of lenition in a
decisively lenitive environment - intervocalic and removed
from the accent - is strong evidence for a fortis-lenis scale.
Apparently, the tendency for such lenition is found in
Mongolian, but it is so minor as to be realized in only the
weakest environments.

The sole (apparently) Altaic language that maintains
the lenition of the internal dorsal tenuis throughout is Korean.
Hence, Lamutic (and Evenkic) tekun 'to become angry'
corresponds with Korean cigindi 'to anger' (55). Indeed, this

in addition to the heavy aspirate variant of the initial position
(compare sections 3.1.1 and 3.1.2) strengthens the hypothesis
that Korean may indeed be an Altaic language.

3.3 The Altaic Media

At least in initial position, the discriminative mark of
the original Altaic media *b, *d, *g may well have been the
lack of aspiration (Poppe 1960:20). Thus, the only opposition
that could occur in initial position - the opposition that is
usually rendered as, for example *b:*p - would phonetically
have been [p]:[p ]. This opposition on the basis of aspiration
without pertinent realization of voice would be typical of an
aspirate fortis-lenis sytem, such as that of Welsh (compare
Griffen 1985:chapter 7).

Unlike the aspirata and tenuis, which appear to have
been to some extent positional variants (the aspirata in initial
position, the tenuis in internal and final position), the media
could appear in all positions. Since the discriminative mark of
the opposition between the aspirata and the media was
aspiration and that of the opposition between aspirata and
tenuis was also aspiration, the differentiating opposition
between tenuis and media must have also been in degree of
aspiration. Thus, the aspirata would have been most aspirated,
the tenuis less aspirated, and the media least aspirated
(unaspirated). This three-way opposition of aspiration most
clearly points to an aspirate system consistent at the outset with
the aspirate fortis-lenis scale of Germano-European.

Once again, Korean deserves particular attention, for
its obstruent system even today maintains three degrees of
aspiration (compare Kim 1970). Assuming that it is an Altaic
language, it would appear to be the most conservative of this
language family. When one looks at the distribution of the
Altaic language family and when one does add Korean to the
family, then the geographic position of Korean on the extreme
periphery and indeed on a peninsula blocking early language
contact on three sides would lend this language the isolation in
which conservativism might flourish (compare Bartoli 1925 -
see section 4, below).

Maintaining the traditional media symbols b, d, g for
this group, the development of the Altaic media in initial
position and in internal and final position is outlined in table 6
(pp. 46, 47).

3.3.1 Labial. In initial position, all of the central group
of Altaic languages maintain the media unchanged. Thus, the
word for 'grey' is realized in Mongolian as boro, in Evenkic as
borog, and in Old Turkish as boz (20).

The only change occurs in Korean, in which the media
is provected in initial position to a tenuis. For example, proto-
Altaic *baragun 'right' is realized in Mongolian barayun

'right', in Goldic (Nanaj) baro/baru 'toward, in the direction
to', and in Korean pari 'to be straight' and paro 'straight,
direct' (21). Such a provection in this position is consistent
with the third expectation for change outlined in section 2.1.
While the change in this degree along the fortis-lenis scale is
not typical in the Germano-European period, it is certainly
found in the High German sound shift (compare Griffen 1981).
In final position, the labial media is subject to lenition
to a susurrata in Mongolian and partially in Tungusan, quite in
keeping with the development of an aspirate fortis-lenis
language and parallel with changes in Germano-European -

MOTHER TONGUE

Issue 22, May 1994

Table 6: Altaic Media
Labial - *b
Initial

proto-Mongolian

Written Mongolian

Manchu

Tungusan

proto-Turkic

Old Turkish

Korean

Internal Strong

proto-Mongolian

Written Mongolian

Manchu

Tungusan

Old Turkish

b 3

Dental - *d

Initial

pre-Mongolian

proto-Mongolian

*d, *3 4

Written Mongolian

d,3 4

Manchu

d,3 4

Goldic

d,3 4

Tungusan

pre-Turkic

proto-Turkic

Old Turkish

i 5

Final

proto-Mongolian
Written Mongolian
Manchu

*w , *b

w/\b 2
b

Tungusan

b/w

Old Turkish

Internal Weak

pre-Mongolian
proto-Mongolian

*y, *w, *j

Written Mongolian

Y,w,j

Manchu

b/w

Goldic

null

Tungusan
Old Turkish

w
b

Final and Internal

pre-Mongolian
proto-Mongolian

Written Mongolian

Manchu

Goldic

Tungusan

Old Turkish
Evenkic, Lamutic
Jakutic
Sagaic

*d
*d

d,3 4
d,3 4
d,3 4
d

d
d
t
z

indeed, with the Indo-European sound shift itself. In
Mongolian, the media is maintained at the end of monosyllabic
verbal stems, but it is reduced to the susurrata *w (*u)
elsewhere. Hence proto- Altaic *ab- 'to take' yields (Middle)
Mongolian ab-, proto-Altaic *abuca 'the taking' yields

Mongolian abuca (44); but proto-Altaic *tabl(g)ai 'have'
yields (Middle) Mongolian taulai or tawlai (44).

The difference between strong position (before a short
vowel) and weak position (before a long vowel) affects the
media as it does the tenuis (compare section 3.2). In the strong
position, the Tungusan media undergoes a lenition much in
keeping with an aspirate fortis-lenis system. Thus, proto-Altaic
*teberi- 'to embrace' yields (Middle) Mongolian teberi- and

Lamutic towel- with the same meaning (45).

More varied and interesting is the development in
Turkic. While Old Turkish maintained the media, colloquial
modern Turkic languages realize a v or a deletion of the
consonant altogether - a clear weakening along the fortis-lenis
scale. Moreover, in some cases, the labial media may be
realized as the dorsal susurrata y or g. For example, Old

Turkish db 'house' is realized in the Shoric and Kuaric dialects
as ug (45).

In the weak position, the Mongolian labial media fell
together with the tenuis. This change must have originally
involved the change from tenuis to media, for the realizations

MOTHER TONGUE

Issue 22, May 1994

Table 6 (continued): The Altaic Media
Dorsal - *g
Initial

Final

pre-Mongolian
proto-Mongolian

Written Mongolian

*g
*g 6 , *g 7

? 6 ,g 7

Tungusan

proto-Turkic
Old Turkish

q 6 ,k 7
q 6 ,k 7

Chuvashic
Korean

x 6 ,k 7
k

proto-Mongolian

Written Mongolian

Middle Mongolian

Manchu
Tungusan

Old Turkish

> I

£>

q 6 ,
k/g

k 7

Y =

Intervocalic Strong

proto-Mongolian
Written Mongolian

Manchu
Tungusan
Old Turkish

*g 6 , *g 7
§ 6 ,g 7

6 ~ 7
Y>g

Intervocalic Weak

proto-Mongolian

Written Mongolian

Y 6 ,f 7

Middle Mongolian

null

Manchu

X b ,x'

Tungusan

Old Turkish

Y 6 ,g 7

1) final of nominal stems and internal 2) final of verbal stems 3) realized as susurrata in modern Turkic languages
4) before */, *i* 5) realized in various Turkic languages as 3, c, s, etc. 6) postvocalic 7) prevocalic 8) before n, /, r

of both Altaic media and Altaic tenuis in this position all
involve susurratae (see section 3.2.1). Thus, the Altaic word for
a type of curds *abarti yields through intermediate *awarca

Mongolian ayarca (48).

In Manchu, the media may weaken to the susurrata w,
but it may be lenited all the way to nonobstruction (null) in
Goldic. In Tungusan, the lenition to w is more consistent.

3.3.2 Dental. The Altaic dental media has gone
through some changes due to position, with variants of palatal
obstructions realized in initial position of syllables with high
front vowels (/ or i) in Mongolian and Manchu languages and
in initial position of all syllables in Turkic languages. Only in
Tungusan has the d been maintained in all positions. Thus,
(Middle) Mongolian del 'mane' corresponds with Old Turkish
jil {jet) 'mane', realized in Chuvashic as silxe (22). Proto-Altaic

*diluga 'temple, bridle' is found in Mongolian asjiluya due to
the high front vowel, but it is realized as dil 'head' in Evenkic
(23).

While there is no provection in initial position, as
there is in the case of the labials, the lack of change is once
again not particularly indicative of anything. Indeed, the
palatalization may well have blocked any provection from the
original dental media.

The same effects of palatalization can be seen in
internal and final position. In Mongolian and Manchu
languages, 3 is realized before the high front vowels / and i.
Thus, while pre-Mongolian *edin 'lord, ruler' is realized
through ejin as in (Middle) Mongolian ejen, it is realized in
Tungusan as Evenkic and Lamutic edi 'husband' (53). In
Turkic languages, it is realized in different ways, such that the
word just cited is rendered in Old Turkish as ddi 'lord', in
Abakan as dzi, and in Kasakhic as if a (53).

The apparent provection of the media to tenuis in
Jakutic is an isolated case and can be treated as exceptional.
Nonetheless, this development may be related to the general
resistence to lenition of the dental media among the Altaic
languages.

3.3.3 Dorsal. Just as *k has given rise to the variants k
and q depending upon its position within the syllable, *g is
realized in Mongolian as g and g. Such variants are of no

concern here, and in manner of articulation (degree along the
fortis-lenis scale) Mongolian maintains the dorsal media in
initial position.

In the Turkic languages, however, the dorsal media
has changed to a tenuis, once again (as in the case of the
dentals) following a pattern typical of an aspirate fortis-lenis

MOTHER TONGUE

Issue 22, May 1994

system. Provection in initial position creates such
correspondences as Mongolian gendun 'masculine' and Old
Turkish kantu 'self (25). In Chuvashic, the variant occurring
after the syllabic vowel (in final obstruction position) is
realized as x, which may well be a further provection in initial
position.

The nasalization that occurs in a Tungusan word when
the following (syllable-final) consonant is n, /, or r is
interesting, but it does not seem to affect the obstruent system
in general. For example, the word for 'to fear' in Mongolian is
gelme-, in Manchu gele-, and in Evenkic rjele (25).

In Korean, word-initial *g is realized consistenly as &,
showing an initial provection in keeping with the Korean
aspirate system. This means that only the two most aspirate,
fortis members of the opposition can occur in initial position.
This is somewhat reminiscent of the restriction prohibiting the
lenis-most consonants from appearing in initial position in
borrowings into Welsh (see Griffen 1975b).

In stem-final position, the media is maintained in
Mongolian and Manchu (according to the data - the table in
Poppe 1960:57 appears to be in error). Hence, proto-Altaic
*bag is realized in Mongolian as bag 'section, group' and

bagca 'bundle' (58). In Turkic, however, it is lenited to 7 when

it occurs after a vowel, as in the Old Turkish equivalent of the
word cited bay 'section of a tribe' (58).

In Tungusan, the media may be provected to a tenuis
in a process of assimilation with the following consonant. For
example, Evenkic degi 'bird' is related with dekte 'feather'
(58). Such assimilations are evidence neither for nor against the
aspirate fortis-lenis nature of the system.

In intervocalic position, the dorsal media undergoes
the same variation as does the labial. In the strong position
(before a short vowel), the Turkic languages undergo an
expected lenition to the dorsal susurrata y when it is in stem-
final position as well. Thus, for example *sogu 'hart' is realized
in Mongolian as sogo but in Solonic as soyo (58). From the

data, a similar process appears to occur in Mongolian as well,
yielding ayu from *agii 'large, broad' (58). Other such

lenitions to susurratae can be seen in such forms as Evenkic
aglan and awlan 'steppe' (58).

The change from g to x in Manchu was probably
achieved through an intermediate y as well (compare Poppe

1960:61). Thus, for example, proto-Altaic *p ogti 'nipple' is
realized in Manchu asfuxu, probably through the intermediate
*Joyil (61). As such, it would have been typical of the lenition
to susurrata expected in an aspirate fortis-lenis system.

The same changes occur in weak position as in strong
with one major exception - an exception that unambiguously
points to the working of the aspirate fortis-lenis scale. In
Mongolian, the media first changed to the susurrata y in

postvocalic stem position, maintaining the media in prevocalic
stem position. By Middle Mongolian, however, the dorsal
disappeared completely only to be realized as y in modern

colloquial Mongolian. The disappearance and reappearance is
probably an illusion created by a shift of standard dialect -just
as the change in dialect between Middle High (Upper) German

and New High (East Central) German created the illusion that
"German" changed such forms as gdn/gen 'to go' to gehen (see
Griffen 1983).

The interesting thing is that in those dialects typifying
Middle Mongolian, the media in weak intervocalic position
lenites first to the susurrata and then to total loss of occlusion.
This pattern is seen in Welsh, particularly in the mutation
system (compare Morris Jones 1913:164-65), and it parallels
the pattern of change in German in which, for instance, Latin
magister 'master' becomes German Meister (compare von
Kienle 1969:111-12). The pattern in Mongolian is typified by
the development (or alternation) from *bogdrla 'to cut through
the throat' to Middle Mongolian bo'orla- and Mongolian
boyorla- (60). Such a development is clearly within an aspirate

fortis-lenis system and parallel with developments in Germano-
European.

Before the high front vowel, Mongolian is also further
lenited toy. For example, the word 'healthy' developed from
pre-Mongolian *sagi'n through intermediate *sayin to (Middle)

Mongolian sajin (61). While this change is not precisely
parallel to those in Germano-European, it is not inconsistent
with them. Indeed, the same type of phenomenon can be found
in Eastern German dialects in which gesund 'healthy' is
realized with an initial glide.

4. Conclusion

From the foregoing investigation, it should be clear
that the occlusive obstruents of the Altaic languages developed
from proto-Altaic along an aspirate fortis-lenis scale. When
changes occurred, they proceeded along the scale in keeping
with the expectations of change within an aspirate fortis-lenis
system. Although these changes were not always parallel with
those in the development of the Germano-European languages,
they were overwhelmingly consistent with them.

Moreover, the fact that the media contrasted only with
the (heavy) aspirata in initial position and only with the tenuis
in intermediate and final positions provides one of the most
dramatic parallels in the relationship between Altaic and
Germano-European. Without any consideration of Nostratic
(much less of Altaic), the following conclusion was drawn for
the ultimate occlusive obstruent system of Germano-European:

On the basis of the evidence that can be evinced from
the Indo-European languages vis-a-vis the Germanic,
it is apparent that the first change to take place in the
system was a slight provection in initial position as
well as other positions of strength that may have
resulted from shifts in stress. In these positions, the
tenuis [t] became more heavily aspirated as the
aspirata [t ] (and likewise for the other positions).
This yielded the three-member aspirate system [d], [t],
[t h ], from lenis to fortis. (Griffen 1988:163-64)

The stage of development for proto-Altaic, then, is
precisely the same as the stage of development for pre-
Germano-European. It is no wonder that the two show such
close parallels in their development, for they have proceeded

MOTHER TONGUE

Issue 22, May 1994

from the same kernel system along the same phonetic
parameters.

Such a close systematic correspondence along with
the multitude of lexical correspondences (see especially
Krippes 1990b) would favor the hypothesis that Altaic and
Germano-European derive from a common ancestor.
Moreover, the developmental stage of the two-member aspirate
opposition in proto-Altaic and the relatively recent
development of the three-member aspirate opposition in proto-
Germano-European would certainly point to the probability
that the diversification of Altaic and Germano-European is
fairly recent (compared with the ages represented by the proto-
languages).

If in fact a genetic relationship exists (and the
likelihood of that possibility seems quite strong given the
agreement of the systematic and lexical evidence), then proto-
Altaic would represent the more conservative branch - the
branch that had not yet undergone the expansion of the two-
member opposition to the three-member opposition. Indeed, an
examination of the known development of Altaic should shed
light on the hypothesized development of proto-Germano-
European, for the development within the same phonetic
parameter from the same base should yield similar patterns of
change. Such is, after all, the precise phonetic nature of Sapir's
concept of linguistic drift (Sapir 1921:chapter 7).

That being the case, the most conservative of all the
languages in the Altaic and Germano-European family groups
in their occlusive obstruent systems would appear to be
Korean, if in fact it is an Altaic language (as may seem more
likely in the light of this current investigation). In Korean, the
three-member opposition of aspiration is most overtly aspirate
in nature, as demonstrated by Kim (1970). And it more
precisely follows the environmental strictures of the proto-
language (although, to be sure, it does deviate from them in a
way characteristic of an early departure from the protosystem).

Furthermore, given the areal linguistic considerations
of Bartoli (1925 - see also Bolinger 1975:chapter 11 for
English terms and a summary), Korean should be the most
conservative from the norm of the isolated area and the norm of
the later area. As pointed out in Griffen 1988:34-35, the
remaining norm of the principal area is problematic anyway,
and the norm of the lateral area would not apply. Should
Korean prove to be an Altaic language, then, its conservative
nature would be a consequence of its geography. As pointed
out by Krippes (1990a), however, care should be taken in
comparing Korean with other languages.

As for the greater development of Nostratic, no further
conclusions can be made at this point. While the fact that both
Altaic and Germano-European maintain an aspirata as the fortis
could be taken to support Starostin (1989) in the contention that
the Nostratic fortis was aspirate (a notion doubted by
Shevoroshkin ms.), two families out of six is not enough at this
point to tip the scales totally in favor of the aspirate, as opposed
to the glottic.

On the other hand, the maintenance of an aspirate
fortis in both Altaic and Germano-European may well indicate
a close relationship between the two within Nostratic. Perhaps,
these two families originally constituted a single dialect with
the aspirata, as opposed to another dialect with the glottalic.

However, the ultimate composition of the Nostratic
relationships depends upon considerably more study. The close
relationship between Altaic and Germano-European at this
point supports only that relationship. How these two tie in with
the other four will depend upon subsequent investigations of
the phonetic bases for those phonological systems.

REFERENCES

Bartoli, Matteo. 1925. Introduzione ala neolinguistica.

Geneva: Olschki.
Bolinger, Dwight. 1975. Aspects of Language, 2nd ed. New

York: Harcourt Brace Jovanovich.
Bomhard, Allan R. 1992. "The Nostratic Macrofamily (with

Special Reference to Indo-European)." Word 43, 61-84.
Bomhard, Allan R., and John C. Kerns, ms. The Nostratic

Macrofamily. Prepublication manuscript.
Campbell, Lyle. 1990. "Indo-European and Uralic Tree

Names." Diachronica 7, 149-80.
Dolgopolsky, A. B. 1964. "Gipoteza drevnejsego rodstva

jazykov severnoj Evrazii (problemy foneticeskix

sootvetsvij)." VII Mezdunarodnyj kongress
antropologiceskix i etnograficeskix nauk. Moscow:

Congress Publications.
Fries, Charles C, and Kenneth L. Pike. 1949. "Coexistent

Phonemic Systems." Language 25, 20-50.
Griffen, Toby D. 1974. "The Development of Welsh

Affricates: A Change through Borrowing." Lingua 34,

149-65.
Griffen, Toby D. 1975a. A New Welsh Consonant Shift:

Description and Implications. Ph.D. diss. Gainesville:

University of Florida. [Ann Arbor: University Microfilms

International.]
Griffen, Toby D. 1975b. "Lenis Initials in Welsh Borrowings."

Language Sciences 36, 6-12.
Griffen, Toby D. 1977. "German [x]." Lingua 43, 375-90.
Griffen, Toby D. 1981a. "German Affricates." Lingua 53, 173-

89.
Griffen, Toby D. 1981b. "The High German Sound Shift:

Phonetic Justification." In: Papers from the 1980 Mid-
America Linguistics Conference, ed. by M. M. T.

Henderson, 199-206. Lawrence: University of Kansas.
Griffen, Toby D. 1982. "Voice-Tension Competition in Greek

and Latin." Forum Linguisticum 6, 202-16.
Griffen, Toby D. 1983. "The Swabian Voiceless Vowel." Word

34, 135-74.
Griffen, Toby D. 1985. Aspects of Dynamic Phonology.

Amsterdam: Benjamins.
Griffen, Toby D. 1988. Germano-European: Breaking the

Sound Law. Carbondale: Southern Illinois University

Press.
Griffen, Toby D. 1989. "Nostratic and Germano-European."

General Linguistics 29, 139-49.
Griffen, Toby D. 1993. "Germano-European and the Phonetic

Plausibility Theory." Word 44, 473-84.
Hjelmslev, Louis. 1970. Language: An Introduction. Trans, by

F. J. Whifield. Madison: University of Wisconsin Press.
Jakobson, Roman. 1972. Selected Writings of Roman

MOTHER TONGUE

Issue 22, May 1994

Jakobson. Vol. 1: Phonological Studies. The Hague:

Mouton. [Rpt. of 1929. "Remarques sur revolution

phonologique du Russe comparee a celle des autres

langues slaves." Travaux du Cercle Linguistique de

Prague 2.]
Jones, Robert Owen. 1971. "Comparative Dialectology."

Studia Celtica 6, 168-74.
Kaiser, Mark, and Vitalij Shevoroshkin. 1988. "Nostratic."

Annual Review of Anthropology 17, 309-29.
Kauffmann, Friedrich. 1890. Geschichte der schwabischen

Mundart. Strassburg: Triibner. [Rpt. 1978. Berlin: De

Gruyter.]
Kim, C.-W. 1970. "A Theory of Aspiration." Phonetica 21,

107-16.
Krippes, Karl. 1990a. "Problems Concerning the Comparison

of Korean with Other Languages." Mother Tongue 10,

article 5.
Krippes, Karl. 1990b. "The Altaic Component of a Nostratic

Dictionary." Mother Tongue 11, article 2.
Miller, Roy Andrew. 1971. Japanese and the Other Altaic

Languages. Chicago: University of Chicago Press.
Morris Jones, John. 1913. A Welsh Grammar: Historical and

Comparative. Oxford.
Poppe, Nikolaus. 1960. Vergleichende Grammatik der

altaischen Sprachen. Wiesbaden: Harrassowitz.
Ramstedt, G. J. 1957. Einfuhrung in die altaischen

Sprachwissenschaft. Bd. I: Lautgeschichte. Ed. by Pentti

Aalto. Helsinki.
Sapir, Edward. 1921. Language: An Introduction to the Study

of Speech. New York: Harcourt, Brace and World.
Scholes, Robert J. 1968. "Phonemic Interference as a

Perceptual Problem." Language and Speech 2, 86-103.
Shevoroshkin, Vitalij. ms. "Review of B[omhard]'s Review of

Typology..." Mother Tongue 10.
Shirokogoroff, Sergei M. 1931. Ethnographical and Linguistic

Aspects of the Ural-Altaic Hypothesis. Tsing Hua Journal

6. [Rpt. 1970. New York: Humanities Press.]
Starostin, S. A. 1989. "Nostratic and Sino-Caucasian." In:

Explorations in Language Macro-Families, ed. by V.

Shevoroshkin, 139ff. Bochum: Brockmeyer.
Trubetzkoy, Nikolai S. 1969. Principles of Phonology. Trans.

by C. A. M. Baltaxe. Berkeley: University of California

Press.
Vladimircov, B. J. 1929. Sravnitel'naja grammatika

mongol'skogo pismennogo jazyka i chalchaskogo

narecija, Vvednie ifonetika. Leningrad.

von Kienle, Richard. 1969. Historische Laut- und Formenlehre
des Deutschen. 2nd ed. Tubingen: Niemeyer.

IE LARYNGEALS — EVER LISTENED
TO THEM?

W. WILFRIED SCHUHMACHER
Rise National Laboratory, Denmark

Modern civilization almost is condemned to create the
NEW. The new, all that differs from what hitherto has been,
has no easy time though. Even experts sometimes are not
capable to see a difference from the insignificant, the over-the-
top, or the valueless. Is the fundamentally new possible at all?
This is denied by F. Fukuyama and others claiming that
postmodernists have witnessed the end of history,
characterizing therefore the question of the new — that central
category dating from the middle of the 19th century — as not
opportune. Contrasting with this view, one can point to a
constant dialectics of devaluation and revaluation, a turning of
the valuable and the valueless, creating new oppositions and
tensions without ever ending up in a synthesis. Nothing
disappears totally but can always enter again the process of
revaluation. Following the logic of innovation, the ground has
to be dug constantly so that what is below comes up, and what
is above goes down. Innovation, therefore, is displacement of
things according to value boundaries.

The past decade has witnessed a revaluation of the
genetic classification of languages and the implications of such
a classification for the prehistory of the human species, thus
digging up again 19th-century scholarship. As for the historical
metalinguistics in the work of Joseph Greenberg, one surely
would be able to set up another metalinguistics — or even let it
be at all! Mutatis mutandis, Ludwig Wittgenstein seems to be
relevant characterizing Sigmund Freud's theory of human
prehistory as pure speculation, something even preceding the
set-up of a hypothesis: Freud's metapsychology is indeed
META-psychology.

Wittgenstein's countryman Ernst Mach (1838-1916)
goes one step further: when being asked about the atoms,
Mach simply answered: "Ever seen them?"

IN THE PUBLIC MEDIA

The following article is reprinted with permission from Science
News (vol. 145 [1994], p. 84), the weekly newsmagazine of
science, copyright 1994 by Science Service, Inc.

SIBERIAN SITE CEDES STONE-AGE
SURPRISE

On a windblown terrace above Siberia's Lena River,
Russian scientists have unearthed evidence that humanity's
evolutionary ancestors inhabited parts of northeastern Asia and
could have made initial forays into North America much earlier
than previously thought.

Preliminary soil analysis by two U.S. geologists

MOTHER TONGUE

Issue 22, May 1994

indicates that stone tools found at the Siberian location, known
as Diring, date to around 500,000 years ago. However, Russian
investigators date the artifacts to at least 2 million years ago,
argued excavation director Yuri A. Mochanov last week in a
talk at the Smithsonian Institution in Washington, D.C.

"I suspect the artifacts are younger than Mochanov's
estimates," says Richard B. Potts, a Smithsonian archaeologist
who examined a dozen stone flakes and blades brought from
the site by Mochanov. "But even if Diring is only 50,000 years
old, it's significantly older than any other human site in
Siberia."

No other human sites in Siberia date to more than
35,000 years ago. This fuels the view that North America's
initial settlers arrived no earlier than 20,000 years ago (SN:
6/9/90, p. 360).

Mochanov, an archaeologist at the Russian Academy
of Sciences in Yakutsk, accepted this theory until shortly after
he started working at Diring in 1982. Geologists digging up
soil samples along the Lena River found some human bones
and alerted Mochanov. He and his coworkers then excavated
several human burials dating to 10,000 years ago and the
35,000-year-old remains of mammoth hunters.

The investigators also found sharp-edged stones that
looked like human tools. These flakes, choppers, and other
implements had been sandblasted by Siberian winds. Only East
African stone tools that date to between 1.8 million and 2.5
million years old resemble the Diring artifacts, Mochanov
contends. The tool-bearing soil has yielded no bones, probably
because they were destroyed by windblown sand, Potts says.

A larger scientific team returned to Diring in 1983.
Annual fieldwork since then has yielded more than 4,000 stone
tools over an area the size of four football fields, making Diring
the largest Stone-Age dig in the world, according to Robson
Bonnichsen, an archaeologist at Oregon State University in
Corvallis, who visited the site in 1992.

Measurements of magnetic reversals and radioactivity
in Diring soil — the latter relying on a technique unknown to
Western scientists — place the finds at 2 million to 3 million
years old, Mochanov maintains.

He offers the radical proposal that direct human
ancestors evolved not in Africa, but in the northernmost
reaches of Siberia, where severe cold forced innovations in
thought and behavior that fostered human evolution.

Archaeologists who have seen the Diring artifacts
generally agree that someone intentionally made them, but they
express skepticism about Mochanov's age estimates. In fact,
thermoluminescence dates for two soil samples collected at
Diring last summer by Michael Waters, a geologist at Texas
A&M University in College Station, place the stone tools at
about 500,000 years old.

Thermoluminescence dating of eight additional soil
samples gathered by Waters will continue. Steven Forman of
Ohio State University in Columbus directs the analysis, which
estimates age from measures of the radioactive signal in sand
grains and the dose of radioactivity in surrounding soil.

Ongoing soil and pollen analysis at Diring will help to
establish whether its inhabitants endured bitter cold or lived
during a relatively warm spell, Oregon State's Bonnichsen
notes.

If cold weather prevailed, the Siberian findings will
put a chill on the widespread opinion that only Neandertals
adapted successfully to frozen climates, Potts asserts. Still, the
species identity of Diring's inhabitants remains unknown.

Diring's estimated age of 500,000 years also supports
theories that people could have migrated to North America
more than 30,000 years ago, adds Smithsonian archaeologist
Dennis Stanford.

Investigators should expand their Siberian search by
launching excavations at 15 recently discovered sites located
near Diring, Potts remarks.

— B. Bower

The following article is reprinted with permission from Science
News (vol. 145 [1994], p. 150), the weekly newsmagazine of
science, copyright 1994 by Science Service, Inc.

ASIAN HOMINIDS MAKE A MUCH
EARLIER ENTRANCE

Members of the human evolutionary family left Africa
and reached eastern Asia 800,000 years earlier than previously
thought, according to a report in the Feb. 25 SCIENCE.

The new estimate comes from a redating of three
Homo erectus specimens from the Indonesian island of Java.
Local collectors found a skullcap of a child at one site in 1936
and partial skulls of two individuals at another site in 1974. The
first skull now dates to about 1.8 million years ago, the latter
specimens to approximately 1.6 million years ago.

Carl C. Swisher III, a geochronologist at the Institute
of Human Origins in Berkeley, Calif., and his colleagues
analyzed the relative proportions of two forms of argon in the
hominid-bearing sediment from the Indonesian sites to
establish new dates for the finds.

Many anthropologists express surprise that H. erectus
ventured to the far reaches of Asia so early. However,
contrasting theories of how the Homo lineage evolved —
which rely on analyses of skeletal anatomy — remain
unchanged in the wake of the new study (SN: 6/20/92, p. 408).

"It shocks me that hominids [members of the human
evolutionary family] lived outside Africa that early," asserts
David W. Frayer of the University of Kansas at Lawrence.
"We'll have to see if these dates hold up."

Frayer and others would prefer age estimates
generated from sediment that still clings to the Indonesian
bones, but Indonesian officials barred the removal of any
material from the fossils, Swisher says. Comparable results at
the two sites, located about 150 miles apart, compellingly
support the new dates, he argues.

H erectus apparently reached Asia before the
appearance of stone choppers and hand axes in Africa around
1.4 million years ago, Swisher holds. This helps explain why
no such artifacts have emerged from Asian sites, he maintains.

African fossils formerly assigned to H erectus may
belong to a separate species that led to modern humans,

MOTHER TONGUE

Issue 22, May 1994

Swisher suggests. In this scenario, championed by Bernard
Wood of the University of Liverpool in England, H. erectus
reached an evolutionary dead end in Asia.

G. Philip Rightmire of the State University of New
York at Binghamton disagrees, based on his anatomical
comparisons of Asian and African fossils.

"Erectus originated in Africa and then pushed out to
Asia in pulses of movement," he argues. "The surprising new
dates indicate that these migrations, and the Homo lineage
itself, have more ancient roots than we thought."

Reasons for the migration of African//, erectus, often
linked to the production of versatile hand-axes, now seem
unclear, Rightmire contends. And much uncertainty surrounds
the relationship of the Homo lineage to African
australopithecines, which consist of the earliest hominid
species. If//, erectus left Africa by 1.8 million years ago, its
ancestor must have evolved simultaneously with various
australopithecines, Rightmire notes.

Milford H. Wolpoff of the University of Michigan in
Ann Arbor offers a third interpretation of the Java dates.
Wolpoff lumps all H erectus fossils into an anatomically
diverse group of// sapiens that evolved in several parts of the
world starting about 2 million years ago.

"These new dates tell us that primitive H sapiens left
Africa much earlier than we thought," he holds.

Recent finds of simple stone tools at a Javanese H
erectus site that may date to 750,000 years ago suggest that
Asian hominids probably concentrated on cutting bamboo with
quickly produced implements, Wolpoff asserts; knowledge of
hand axes may simply not have been put to use.

A related report, published in the March 3 NATURE,
concludes that H erectus and H. sapiens may have lived
simultaneously in China for a short time.

Measures of the rate of uranium decay in animal teeth
uncovered last year in the same deposit as a H. sapiens skull
place the finds at a minimum of 200,000 years old, assert Chen
Tiemei of Peking University in Beijing, China, and his
coworkers. Some Chinese H erectus remains date to 300,000
years old or less.

— B. Bower

The following article is reprinted with permission from Science
News (vol. 145 [2 April 1994], p. 212,), the weekly
newsmagazine of science, copyright 1994 by Science Service,
Inc.

FOSSILS PUT NEW FACE ON LUCY'S
SPECIES

Investigators have recovered and pieced together the
first nearly complete skull of the earliest known species in the
human evolutionary family, Australopithecus afarensis.
Fragments of the 3 -million-year-old skull, as well as of a
number of limbs and jawbones, turned up at the Hadar site in
Ethiopia. Fieldwork at Hadar in the 1970s yielded the first A.

afarensis remains, including the partial female skeleton known
as Lucy. William H. Kimbel and Donald C. Johanson, both
anthropologists at the Institute of Human Origins in Berkeley,
Calif., and Yoel Rak, an anatomist at Tel Aviv University
describe the new finds in the March 3 1 NATURE.

"There is no obvious sign of evolution in this
prehuman species for about 1 million years," Kimbel says. "Yet
later, in only a fraction of that time, [A. afarensis] gave rise to
a great branching of the family tree."

The age estimate for the new skull, assembled from
more than 200 fragments found in a sandy gully in March
1992, makes it the youngest known example of Lucy's kind.
Analysis of two forms of argon in crystals of volcanic rock just
above and below where the skull lay established its age. The
fossil's anatomy confirms that a 3. 9-million-y ear-old cranial
fragment previously found at another Ethiopian site also
belonged to A. afarensis, the scientists argue.

Lucy herself lived about 3.2 million years ago.

Since 1990, annual fieldwork at Hadar conducted by
Kimbel and his coworkers has yielded 53 A. afarensis
specimens.

Kimbel attributes the Hadar skull to a male much
larger than the diminutive Lucy, who stood about 3V4 feet tall.
The skull and other new Hadar material indicate that A.
afarensis males were considerably larger than females,
although average size differences between the sexes remain
unclear, Kimbel notes.

A minority of researchers places smaller and larger A.
afarensis in separate species. However, later australopithecines
— including a lineage that died out 1 million years ago — also
featured large size differences between sexes of the same
species, writes Leslie C. Aiello, an anthropologist at University
College in London in a comment accompanying the new report.

Reconstruction of the three-quarters complete Hadar
cranium will yield an estimate of brain size, Kimbel notes.

A virtually complete ulna, or forearm bone, and a
partial upper-arm bone found at Hadar also came from A.
afarensis, the scientists say.

The ulna curves and is long relative to the upper arm,
a pattern observed in chimpanzees. But the A. afarensis ulna
lacks elbow features that allow chimps to support their body
weight with the forelimbs while walking, Aiello says.

The thick upper-arm bone contains deep grooves
where muscles attached, much like a corresponding A.
afarensis fossil found at a nearby site (SN: 1 1/20/93 p. 324),
Kimbel points out.

Aiello calls the arm bones "ideally suited to a creature
which climbed in the trees but also walked on two legs when on
the ground."

Some researchers argue that female A. afarensis
favored tree climbing, as indicated by their curving toe bones.

Kimbel suspects debate about how Lucy's kind moved
about will continue. A. afarensis clearly could walk efficiently
on two legs, "but I don't know if we can say whether they spent
time in the trees," he adds.

— B. Bower

MOTHER TONGUE

Issue 22, May 1994

The following article is reprinted with permission from Science
News (vol. 145 [2 April 1994], p. 215), the weekly
newsmagazine of science, copyright 1994 by Science Service,
Inc.

MODERN HUMANS LINKED TO
SINGLE ORIGIN

A new study that calculates the mathematical fit of
competing explanations of human evolution with the
geographic array of specific fossil features supports a single
African or southwest Asian origin for modern humans.

The analysis enters a heated debate over human
origins. One theory posits an African genesis for modern
humans between 100,000 and 200,000 years ago, after which
Homo sapiens spread elsewhere and replaced Neandertals. An
opposing view argues that modern humans evolved
simultaneously in several parts of the world beginning about 1
million years ago, with genetic input from Neandertals (SN:
9/25/93, p. 196).

"Africa and southwest Asia are good candidates for
areas where modern human anatomy originated f asserts Diane
M. Waddle, an anthropologist at Duke University Medical
Center in Durham, N.C. "I'm confident that Neandertals had
nothing significant to do with modern human evolution."

Waddle's study relies on a method, developed by
geneticist Robert R. Sokal of the State University of New York
at Stony Brook, for calculating the correspondence between
various scientific predictions and sets of relevant data. Sokal
and his coworkers have used this method to evaluate theories of
modern language origins based on links between language
patterns and genetic traits in European populations (SN:
8/22/92, p. 117).

The Duke scientist studied 83 fossil craniums of H.
sapiens and Neandertals found at sites in Europe, southwest
Asia, and Africa. Specimens ranged in age from around 40,000
to 400,000 years old. Waddle placed the fossils in 12 groups,
depending on geographic location and age.

She then measured a series of cranial features and
traits to calculate the anatomical variation in each group and the
degree to which pairs of groups resembled one another.

Considered either as separate origin sites or lumped
into a single group, African and southwest Asian fossils
account better for the resulting pattern of anatomical
relationships than the assumption that evolution took
distinctive paths in Africa, Asia, and Europe, Waddle contends
in the March 3 1 Nature.

Advocates of multiregional human evolution, such as
Alan G. Thorne of Australian National University in Canberra,
doubt that Waddle's conclusion will hold up once she studies
East Asian and Australian fossils. Waddle plans to analyze
these specimens, which have often been cited in defense of a
separate Asian evolution of modern humans.

— B. Bower

MAMMOTH TRUMPET REPORTS THE
DEBATE ABOUT AMERIND DATES

HAROLD C. FLEMING
Gloucester, MA

Our friendly companion newsletter/journal, Mammoth
Trumpet, had a particularly rich recent issue, focused primarily
on the question of questions for American archeology. While
the editor, Robson Bonnichsen, would let us reproduce the
pictures and whole articles of that issue, it is beyond us. We've
too much already. We do recommend once again that you-all
subscribe to the other MT — Mammoth Trumpet (at Oregon
State University). Or a large strumpet. We will skip the parts
about the Siberian digs and hypotheses of the "mad Russian"
(as some say), Professor Mochanov. Except to repeat what we
reported earlier — he finds Oldowan tools in eastern Siberia at
several my a and thinks humanity evolved in Siberia — , there
is only one thing to add. Mochanov was worried about his
dates, so he gave some of his material to an American lab
which allowed that the stuff was maybe 500,000 years old, not
several million. But if fairly early man lived in the ice box, and
so near to Beringia, then that is a noteworthy discovery! Ah,
but it depends on the tools! Can permafrost break up stones?
When is a broken stone a broken stone and when is it
Oldowan? (= from Olduvai Gorge).

That kind of question dominated the report on the
great conference in Brazil on the validity and age of the Pedra
Furada site. Was it really occupied more than 30,000 years ago
by modern people? Were the admittedly crude stone tools
really made by man? Among others, Scotty MacNeish, Tom
Dillehay, Dina Dencause, and various professional skeptics and
some European experts on stone were at the conference. They
all could agree on very little.

The key questions for all the older-than-Clovis sites
are: (1) How old are the various strata (layers of debris and
earth)? (2) Is the site properly stratified, and have materials
from one age intruded into the period of other materials? (3)
Above all, are the stone, bone, or other materials made by man
or are they products of natural processes? A geologist would
care about God's stone work. We in human prehistory care
about the stones only as evidence of what Bar-Yosef wants to
call human behavior. One could excavate acres and acres of
properly stratified stones and bones and shells and lord-knows-
what — and get each stratum precisely dated and each object
carefully plotted in relation to all the others. You could get a
computer to arrange all this stuff in a kind of 3-dimensional
view of the whole site so that you could see what things lay
above or below or next to what other things.

But the whole bloody picture could be meaningless
culturally!

As the Mammoth Trumpet report put it, the key
question was: Are these stones geofacts or are they artifacts?

Did the action of common processes of nature (a la
Lyell) produce these shaped stones? Or did the actions of
humans produce them?

In Ethiopia, volcanic Ethiopia, one can find pieces of
obsidian ^alc'i/ almost anywhere one walks. One can pick up

MOTHER TONGUE

Issue 22, May 1994

a piece and drop it or throw it so as to break it. Since it is a lot
like glass, one can quite by accident produce a razor-sharp
piece of obsidian to use as a cutting tool. Nature can produce
the same. So just a sharp edge does not necessarily imply an
artifact. Unless there is some evidence of human work to get
that edge or to "haft" it somehow, that sharp /balc'i/ is just a
geofact. (To "haft" something means somehow to attach it to a
handle or device for holding it).

Not only Guedon's site in northeastern Brazil but also
Tom Dillehay's site in Chile (Monte Verde) at lower strata and
also Scotty MacNeish's Pendejo Cave at lower strata — they
each and all share the geofact versus artifact problem. The
archeological skeptics are quoted as saying: Look, it's hard to
tell the difference! If two of us cannot agree that this piece of
stone is an artifact, then we are bound to conclude that it is
either a geofact or too hard to tell, an unclassified piece of
stone. Just a hunk of rock that happens to be in a stratum
30,000 years old!

OBITUARIES

HAROLD C. FLEMING
Gloucester, MA

All those who have escaped the Grim Reaper's wicked scythe
can share the general sadness that we have to report. Three of
our esteemed colleagues have departed for places unknown.
We will miss them all, not only for themselves but also because
their leaving diminishes us as a whole. Each was valuable. But
also two of them were old friends of a quarter century or more.

interdisciplinary approach also created a new field,
archeomythology.

Although skepticism about her ideas was widespread
among scholars, they were welcomed by many feminists and
by Joseph Campbell, the mythologist. Writing about The
Language of the Goddess, Gerda Lerner, a historian at the
University of Wisconsin, said that although Dr. Gimbutas's
theory could never be proved, it could "challenge, inspire, and
fascinate" simply by providing an imaginative alternative to
male-centered explanations.

She was the main proponent of the view that a warlike
people speaking the language ancestral to all Indo-European
languages conquered Europe and imposed their speech and
culture on the conquered. A contrary theory, advanced by
Colin Renfrew, holds that early farmers moving out of the
Middle East across Europe transmitted a common language
along with agricultural techniques.

A native of Vilnius, Lithuania, she received degrees
from Vilnius University and a doctorate in archeology in 1946
from Tubingen University in Germany. She immigrated to the
United States in 1949, conducted post-graduate research at
Harvard University and in 1955 was elected research fellow of
the Peabody Museum at Harvard.

Dr. Gimbutas joined the U.C.L.A. faculty in 1963 and
served as professor of European archeology until her retirement
four years ago. She directed five major archeological
excavations in southeastern Europe.

Dr. Gimbutas is survived by three daughters, Zivile
Gimbutas and Rasa Julie Thies, both of Los Angeles, and
Danuta Lake of Anacortes, Washington.

(Adapted from the New York Times, 4 February 1994)

MARIJA GIMBUTAS

Dr. Marija Gimbutas, author, archeologist, and one of
our great experts on Indo-European prehistory, died from
cancer on Wednesday, 2 February 1994, at U.C.L.A. Hospital
in Los Angeles. She was 73 and had lived in Los Angeles.

Dr. Gimbutas was the author of 20 books and more
than 200 articles on European prehistory and folklore. She was
an authority on the Prehistoric incursions of Indo-European
speaking people into Europe and how they changed society
there.

Three of her more noteworthy books are Goddesses
and Gods of Old Europe (1974), The Language of the Goddess
(1989), and The Civilization of the Goddess (1991).
Collectively, they present an interpretation of the neolithic
period of Europe that challenged traditional views of
prehistoric societies.

Perhaps her most controversial proposal was that the
world was at peace during the Stone Age, when goddesses were
worshipped and societies were centered on women. Then, the
theory went, about 6,000 years ago, a European culture in
which the two sexes lived in harmony was shattered by
patriarchal invaders, and the worship of life-giving goddesses
was replaced by reverence for war-like gods. Her studies and

SHERWIN J. FEINHANDLER

Sherwin Feinhandler died early this year at 58. He
used to be amused by people who thought his label,
Feinhandler, S. J., meant that he was a Jesuit priest and sought
to confess to him. Others did on occasion confess to him
because clinical psychology was one thing he did during some
periods of his career.

Sherwin was a multi-faceted human being. The full
roster of his professional activities can be read in the newsletter
of the American Anthropological Association, written by John
Mason and Tom Harford of Washington, D.C., plus myself.
Suffice it to say for our purposes that he was at once an
anthropologist, linguist, psychologist, businessman, teacher,
and consultant. He would not be pinned down to one
profession, nor one interest. He scorned specialization,
although he was able to become awesomely well informed on
a variety of topics (e.g., automobiles, computers, statistics, fine
wines, drugs and their properties, etc.).

A graduate of Northwestern (BA) and Syracuse (MA),
he got to Harvard just as the Chomskyite revolution was hitting
the Boston area, especially in psychology. I can vouch for the
excitement in the academic air those days. Riding the wave of

MOTHER TONGUE

Issue 22, May 1994

theoretical and interdisciplinary dynamics generated by that
revolution, he gained a mental perspective on three fields
which he never lost.

In anthropology he did his stint of field work, working
with the Kamba of Kenya and among other things gaining
fluency in Swahili. After his field notes were stolen, he roamed
around East Africa for a while despairing of his doctorate. He
ran a small business in the former Belgian Congo during the
Lumumba period and watched UN operations with fascination,
especially the cruelty of Ethiopian soldiers to the Congolese.
Finally, regaining his morale, he returned to Harvard and got
his Ph.D. But his lost field notes had cost him his credentials as
Africanist and so he became officially a psychologist cum
social science "type", not the Africanist he had wanted to be.

During his misfortunes in Africa, he observed that
people often failed to treat him with proper respect but that
upper class Englishmen lacked that particular problem. Coming
from an upper bourgeois German Jewish background in
Chicago, secure and treated well by the world, he had been
shocked deeply by the great Holocaust, by the lingering albeit
mild anti-Semitism of much of academia, and his precipitate
drop in status in Africa. How he reacted to all this produced one
spectacular theater of role-playing which lasted the rest of his
life and became so habitual and embedded that his final self
was the role.

Sherwin decided that he would become an upper class
Englishman! This was preferable to being "a little Jew from
Chicago", as he saw it. He worked on his RP (received
pronunciation or aristocratic English) with his sharp linguist's
ear; he married a beautiful young Englishwoman (blonde and
blue of eye) and spent lots of time learning the culture of the
British upper strata and then living that culture as his own. His
efforts were doubtless helped by his good German Jewish
genes. He was almost tall, lean and fit, blonde, blue-eyed and
nearly as handsome as a movie star. Some were reminded of
Ronald Coleman. And he was terribly smart; I've hardly met
anyone brighter.

To reflect on his role-playing. What Sherwin managed
to do was to hold the customary role (status) of an American
social scientist, while playing it as if he were really Sir
Laurence Olivier. He confused many people in academia and
— such is the social insecurity of many people — he infuriated
others who thought he was "putting on airs" or "talking down
to them". After all, he was an American with a name like
Feinhandler! And if he "really was only a little Jew from
Chicago", how dare he speak with a British accent as if to raise
himself above or belittle his listeners? And they punished him
for it.

We need a bit of socio-linguistics. {Mother Tongue
has obviously neglected that sub-field; it seems irrelevant to
long range issues.) Foreigners may not know these details, just
as Americans often fail to hear the nuances in the conversations
of others. In the English-speaking world of America the dialect
with the highest prestige is not a native dialect; it is England's
RP. Americans look upwards socially to aristocratic Britain.
With due respect to William Labov, whose fine work on
American speech is saluted, he has missed this point. What
shows up most in some old dialect areas around New York and
Boston is imitation of RP of various periods. In 20th century

America, a widespread form of this imitative RP is often called
a "pseudo British accent". Perhaps the best example of this
dialect is offered by the conservative William F. Buckley, Jr,
who mastered an aristocratic drawl but not British vowels, or
actresses like Lauren Bacall (contrasting markedly with
Humphrey Bogart's more ordinary American). One can hear
this Elegant English frequently on the East Coast, on campuses
(English departments especially), in art circles, and among
parvenus, the nouveau riche, etc.

Pseudo-British occurs internationally too. Some of us
once heard in Mogadishu (Somalia) the voice of Bill Buckley
coming from a Indian teaching English literature at Somali
National University. An astounding replica of America's most
famous pseudo-Brit!

Sherwin's great accomplishment was to master RP to
the extent of getting it all but the localisms. Bill Buckley has
fooled no Englishmen probably but Sherwin did. Several times
I have heard Englishmen (in Boston pubs) ask Sherwin where
he came from in England because they could not quite spot
which exact local area it would be. He usually replied:
"London." That great metropolitan place could sprout dialects
as individual as his.

Since the clinician and participant predominated in
him, we have few publications to attest to his theoretical
prowess.

Sherwin became and remained for years a compleat
Bostonian, Cambridge variety, but one with a British flavor.
Most loyal of friends, good company and a powerful intellect.
Oh, he is missed!

Postscript: Sherwin would be amused at the irony of
his own case. For most of the past decade he ran and ran, until
he developed into near marathon fitness. Most days he ran 10
km and he won a few races of that distance in his age group. He
got so fit that his cholesterol level was normally under 160 and
his body showed no fat. He even quit smoking. All this irritated
the Grim Reaper so much s/he gave Sherwin leukemia just for
spite.

SUSAN PARK

Susan Park, formerly Mrs. Willard Z. Park, died of
cancer at age 85 last October. Unfortunately, we were not
informed until February this year because her family was
severely grief-stricken and did not wish to accept her demise.
A very sad state of affairs.

But as a consequence of that we lack much of the vital
data on her earlier life. What we do know is that she came from
a wealthy upper class family in San Francisco. They were
German Jews in origin but seemed to have been spared the
experiences with anti-Semitism that shaped the lives of others.
Ethnicity played virtually no role in Susan's life — but politics
did.

Susan studied anthropology at Yale in the 1930s,
especially with Edward Sapir and George Peter Murdock. She
also knew (as a peer) Morris Swadesh, whom she loathed. Her
professional career as an anthropologist ended in the 1930s
because she married another graduate student anthropologist,

MOTHER TONGUE

Issue 22, May 1994

Willard Z. Park. Being wife and mother consumed most of her
time for the next few decades, while Willard became an
important person in Latin American affairs for the U.S.
government and, later on after the end of World War II,
UNRRA representative in Ethiopia. Becoming deeply involved
with reconstructing Ethiopia after the war, Susan and Willard
fell in love with that country and stayed on by hook or by crook
for the rest of their lives, or as long as they could.

Willard's career which had been flying at the
ministerial level for some years was abruptly and brutally
smashed by the House Un-American Activities Committee
(U.S. Congress) in the 1950s. That unfortunate committee, a
consequence of the Cold War, was not able to destroy the lives
of everyone suspected of being a Communist; a few survived.
Willard left the encounter no longer an employable
anthropologist or bureaucrat; he was finished in those realms,
even though, as he swore for years ever after, he had just been
sympathetic to the plight of the downtrodden and had never
been a member of the Communist party. Fellow Yale Man
Morris Swadesh suffered the same fate, as is well known.

Picking up the pieces of their lives, the Parks moved
back to Ethiopia and cobbled together an adjustment, partly
export and import business and partly consultant work for
government. Willard succumbed to a penicillin allergy in the
1960s. Susan eventually returned to the USA after the
Ethiopian Revolution and lived in Nevada working part time on
her Paiute ethnography and book. (She and Willard had first
done "his" obligatory field work among the Paiute of the Great
Basin [western USA].)

However, such a bare bones account of the facts of
their lives neglects the considerable help they gave to the cause
of anthropology in Ethiopia. Their home was a meeting place,
their government contacts were numerous, their help
unstinting. They saved my first field trip by helping me get a
local job after my grant money ran out sooner than expected.
They helped many others but also poked into many aspects of
Ethiopia's numerous cultures as is the wont of their still beloved
anthropology. What they did not do, sadly, was publish their
observations, thoughts and conclusions. Susan continued
poking around Ethiopia and talking with their very many
Ethiopian friends for years after Willard died.

Having to do the marriage and family bit like a proper
woman of her day, even though this cost her the doctorate in
anthropology, Susan stayed faithful to the discipline that she
traveled so far from her sheltered and privileged home in San
Francisco to study. But the good she did for others was
surpassing. It is still deeply appreciated.

QUICK NOTES AND HINTS OF
THINGS TO COME

HAROLD C. FLEMING
Gloucester, MA

Quick Notes is designed to fill in the small holes of this issue.
It is an awfully good way to present proposals in a nascent form

(perhaps a fetal form would be more apt) and see what
reactions they get. No one is responsible for bibliography or
"proofs" or formal discussion in Quick Notes. Hunches reign
here. It is also a good place to present small bits of news of
interest to people. Let the italics present and define the topic.

Gerard Diffloth is back in touch, but just for a moment. Later
in this spring, he is off to Cambodia for a year's sabbatical to
search out the many small languages still undescribed and not
yet assessed taxonomically. They are hiding in the jungles and
other places. Some of that clearly will be dangerous to Gerard's
health, but he says the mosquitoes are more to be feared than
Pol Pot. Let us hope our good colleague brings back a Bolyu or
Ongota or two!

Same fellow has to disagree, but amicably, about the
classification of Bolyu. His hunch is that Bolyu (Lai) belongs
in North Mon-Khmer near what used to be the Palaung-Wa
cluster.

Same fellow has a "far out" hunch, nothing definite at the
moment, that the Munda group of languages (in India) may be
farther away from the rest of Austroasiatic than experts
presently think. Munda presently is roughly the western half of
Austroasiatic, or a sub-phylum. But maybe it is a proper
phylum in its own right. For some reasons. Such a change
would give India four native phyla (Burushaski, Nahali,
Kusunda, Munda) plus its very old intruders (Dravidian,
Tibeto-Burman) plus Indie.

John Colarusso has joined ASLIP, so as not to miss the uproar
over his I-E is related to PNWC hypothesis. John will take the
criticisms and respond vigorously but in his usual amiable way.
John was been looking over Dene-Caucasic on his own and
does think that Sino-Tibetan has a good chance of being related
to Na-Dene (as Sapir, Shafer, and Pinnow thought). One wants
to point out to long rangers that such an agreement is no small
thing.

Kenneth Jacobs has joined ASLIP too, so our Canadian
contingent is now quite robust! Ken is setting up a session at
the American Anthropology Association meetings in Atlanta
(late November) which will focus strongly on the prehistory of
I-E. Three of our long rangers will be on the panel. Look
forward to attending it!

Three of us in various discussions have concluded that there is
one thing about the Chomskyite and post-Chomskyite period in
linguistics that we dislike intensely. With all the emphasis on
theory — nay, the tremendous over-emphasis on theory —
most of the work of gathering the data on the whole human
inventory of languages is being done by a minority of scholars.
For some, like Diffloth, there is considerable personal danger
involved. For others just hard work with precious little
academic reward, because rewards seem to gravitate to
theoretical hot shots. We field workers could use some help!
Languages are dying out! Some of you need to back away from
your computers and do field work\ The experience will give
you a "feel" for the languages you deal with, above and beyond

MOTHER TONGUE

Issue 22, May 1994

what you get from reading the publications of others.

Victor Mair has been written up in Discover magazine. I heard
him also on National Public Radio (USA); he's been seen on
TV. Those Sinkiang mummies are just so intrinsically
interesting that Victor's fame is natural. As far as the I-E impact
on early China is concerned, Victor can easily get into hot
water with the modern Chinese. What Victor has said already
may yet cost him dearly in terms of research permits — the
Chinese officially at least do not wish to hear it. Such is an old
problem in ethnography, as everyone knows.

Edwin Pulleyblank has been stressing I-E and Chinese
connections for decades now. Maybe 30 years ago he discussed
the Tocharians in this context and said essentially what Mair
has now said. Victor in fact says that he hopes that everyone
realizes that his report was not a formal publication because he
did not list any bibliography. Had he done so we would have
confronted a considerable literature on Tocharian and/or
Iranians vis-a-vis China in a number of languages. Indeed
nearly a half century before Pulleyblank's writings and 77 years
before Victor's a German scholar had said very similar things.
Plus 9a change...

Tocharian or Iranian? A debate is shaping up about the
language spoken by the Caucasoid mummies in Chinese
Turkestan (Sinkiang). Good people on both sides. Good
arguments on both sides. One says it was early Iranian or Indo-
Iranian that was spoken in Sinkiang; the other says no, it was
Tocharian. We really do not know, of course, but one can
suggest that people be pressed to stipulate what reasons they
have for choosing one alternative. Some of the reasoning so far
has been rather sloppy.

Peer pressure. Just to show everyone how detached and
scientific our heads are: consider this example. One colleague
reports he is under considerable pressure from another
colleague to "decide" that the mummies "must" have spoken
Iranian. If someone asks him then, "why do you think the
mummies spoke Iranian", he can reply: "because my colleague
professor X obliged me to believe that!"

Another fossil companion for Lucy. The NY Times reported
that new fossil finds in the Danakil (Afar lowlands) enlarged
the range of characteristics of the Australopithecines, still 3.5-
4.0 mya. (See p. 52 of this issue for the report in Science News.)

Ofer Bar-Yosef 'has much to tell us but no time to do it. He will
give us a fuller report of his excellent activities in May when
things cool down for him. We'll report on it in the Summer
issue. We are pleased to report that Ofer stood for election to
the Board of Directors of ASLIP (in absentia April 16th). He
won.

Scotty MacNeish is a hard man to beat down. Now he has gone
to an new site in New Mexico, there to continue his efforts to
convince the skeptics. It probably would be easier to charm
them or buy them off. Scotty will have the absolute world's
record for digs!

The Journal of Afro-Asiatic Languages has risen like Phoenix
from its supposed ashes, under the inspired direction of old
friend, Ephraim Isaac (Princeton, NJ). This is deeply pleasing
to many of us Afrasianists. A full field report on Ongota, sans
taxonomy, will probably come out in two months. Rush your
subscription orders to Ephraim Isaac, Institute of Semitic
Studies, 195 Nassau Street, Princeton, NJ 08542, USA.

It is not too late to get on the programme for the World
Archeological Congress, Section 3, Language, Anthropology
and Archeology to be held in New Delhi, India, December 4-
11, 1994. At this point, it's probably a good idea to send the title
of your proposed paper plus a brief summary to Dr. Makkhan
Lai, World Archeological Congress, P.O. Box 112, H.P.O.,
Aligarh - 202001, India. Phone (571) 29143, Fax (11)
6862049, Telex 564-230-AMU-IN. For electronic high-
waymen, try Prof. V. N. Misra, Deccan College, Post Graduate
& Research Institute, Pune - 411 006, India whose E.Mail
address is dc ! mishra @ eucca. ernet. in. Their second
announcement lists a fair number of our long rangers with,
however, fairly low participation by Americans and
Americanists.

It is also not too late for Africanists to plead to be included in
the program of the International Conference on Trends in the
Historical Study of African Languages at the University of
Hamburg (Germany) on 5-7 September 1994. Interested parties
should write to our very own Ekkehard Wolff or other
organizers, Ludwig Gerhardt and H. Meyer-Bahlburg. Looks
like some quite interesting papers by good solid people will be
presented! Go!

For his many admirers we give the new address of our
esteemed colleague Mr gen Pinnow (he has dropped the Hans).
On his wee island in the North Sea, his address is: Gorch-Fock-
StraBe 26, D-25980 Westerland/Sylt, Germany. Write to him.
He knows a lot!

Paul Benedict wants us to correct our misstatement of his belief
about Bolyu. In MT-21, page 42, we say that Bolyu is "at best"
a Mon-Khmer language. He thinks that makes no sense and
wants long rangers to know that he said that Bolyu is "at heart"
Mon-Khmer. On the question of Southeast Asia as a "nesting
area", Paul thinks it was indeed for Austro-Thai, et al., and with
"PST speakers on western fringe, early movements to south
and west apart from Han to northeast. Movement from west
unlikely. Best to consider native also — I see no good evidence
for the tie with Caucasic et al. — but John (Bengtson) and I are
supposed" to argue this point.

Same fellow accepts Hal's comparison of himself with
Colarusso as much alike, although Austro-Tai is much more
complex than his IE. But "I'll argue with him, however, about
his view that what he calls 'Morphological cognates' are any
more significant, or better maintained, than what he calls
'Conventional cognates'. In the Japanese/ Austro-Tai book, I
present over a dozen pages (123-136) on the former, here
affixes of all sorts, as compared with over a hundred (161-264)

MOTHER TONGUE

Issue 22, May 1994

on the latter!" He has many things to say about reconstruction
but we will save them for MT-23.

Andrew Merriwether of U/Pittsburgh, and an up-and-coming
star in the biogenetics heaven, gave a talk on mtDNA in the
Americas. He will publish much of it elsewhere but he has
promised to give us a summary report in ordinary English soon.
His study is the most comprehensive study of native Americans
from the Arctic Eskimo to the Amerinds of Tierra del Fuego.
He will give us a cladogram of quite a few "tribes" which we
will match up against the Greenberg Amerind taxa viewed as a
matching cladogram. It should be very interesting! One of his
results, which is given here in a very preliminary form, is his
conclusion that Eskimoans, Athapaskans and Amerind all go
back to a single founding lineage before they join the rest of the
world. I.e., his mtDNA findings do not confirm Torroni's
findings or those of Turner and Segura which separate
Eskimoan, Na-Dene, and Amerind peoples. At the moment, he
seems to be saying that the three native American stocks are
more like each other than they are like Asians but that the
whole lot relates to eastern Asians (not necessarily Siberians)
more than to the rest of humanity. Andy is very solid, and I
urge people not to dismiss his findings out of pure frustration,
not just yet anyway, until he has presented his full data and
conclusions.

LETTERS TO THE EDITOR

7 March 1994

Dear Hal:

I enclose my check for membership in ASLIP. Mother
Tongue is wonderful. I appreciate the copy you sent, and I
enjoyed your comments on our paper a lot. I want to discuss
some of the issues that your raised. I don't disagree with any of
them, but a few points deserve more elaboration.

I am sorry that our paper seemed so technically
complex. (I ought to point out that linguistics is forbidding to
most of us and that there is an element of the pot's argument
with the kettle here.) But in both of our fields, the technicality
is necessary, and things will probably get worse rather than
better in the future. You also remarked that we ignored
mitochondrial "Eve", which is correct. I will write you a few
paragraphs about both of these points, hoping to convince you
that the genetic theory is not so bad after all, to explain why we
ignore Eve, and to illustrate the perils of proceeding without the
theory, under the assumption that data speak for themselves.

For three years after the papers about Eve appeared, I
assumed that there was something that everyone but me
understood and that my skepticism about the whole thing
reflected some inadequacy on my part. Then I looked into it,
talked with population geneticists, and realized that I was not
alone in thinking that the estimated date of Eve was irrelevant
to issues of modern human origins. Population genetics theory
is gobbledygook not only to linguists but to many (most?)
molecular geneticists, and the fuss about Eve was the result of
doing genetics without genetic theory.

Here is a simple thought experiment that I use in my
classes. There is an island with a population of exactly 100
humans on it, perhaps an experimental colony run by aliens.
Every generation there are exactly 50 males and 50 females,
and each female has exactly two children who survive. This
island has been going for a very long time. So this is a
population with no history, no dynamics, and no selection.
Every single individual contributes an equal amount of genetic
material to the next generation. Let us think about the mtDNA
on this island.

Each woman has exactly two children each
generation. About a quarter of the women have two daughters,
a quarter have two sons, and half the women have one child of
each sex. But we are concerned only with pedigrees of females
through females when we study mtDNA, and this female-only
pedigree is not so dull after all.

The fifty women on this island have roughly 37
mothers since the other 13 women last generation had two sons.
So fifty women have 37 mothers. Going back another
generation, our 50 women have about 27 mothers' mothers.
Every generation the number of female ancestors through
female-only lines decreases by about %. Before long, there is
only one female at the root of the mtDNA tree, the
mitochondrial Eve of this island population. Probability theory
suggests that, on average, this island Eve lived 100 generations
ago, say 2500 years assuming 25-year generations. (The
expected depth of the tree in a population with N breeding
females is 2N generations.)

What happened 2500 years ago? Absolutely nothing!
This population has no dynamics, no history, no "origin",
nothing at all.

So we get a grant to study mtDNA on the island, go
through a mysterious computer analysis of the data, and derive
an estimate that Eve lived 2500 years ago. What does this tell
us? What it tells us (with a large standard error) is that the
number of women on the island is 50. It doesn't tell us anything
at all about origins or about history.

All of this is intuitive to population geneticists
because, I suppose, they are trained to think this way. So what
does the 200,000 year estimated root of our species' mtDNA
tree suggest? It suggests that our species has 5000 breeding
females each generation. (Mitochondrial DNA also suggests
that there are 50,000 breeding female chimpanzees.) I have
never been able to understand why the 200,000 year date was
seen as support for the Garden of Eden model of human origins
as opposed to the multiregional model. I think it is nearly
irrelevant to the issue. As I said above, I assumed that I was
missing something for three years, now I think that everyone
else missed something!

A true fact from population genetics is that the speed
of neutral change is inversely proportional to population size,
so if a population shrinks it loses variability very fast because
it is small. If it grows again to a large size, it recovers
variability very slowly because it is big. Hence the estimate of
5,000 human breeding females is thought by many to reflect a
rather severe bottleneck in our history. To be honest, very few
of us would take this evidence from one locus very seriously if
it were not concordant with evidence from other systems. The
bottleneck in human history has been known about and widely

MOTHER TONGUE

Issue 22, May 1994

accepted by geneticists since the 1970s. What we tried to do in
the Current Anthropology paper was to estimate when it
occurred or, better, when the recovery began. If we believe
current mutation rate estimates, it seems to have started around
50,000 years ago with a big standard error.

Here are a few other comments about your comments.
It seems as if the ancestors of Africans were the first to undergo
the late Pleistocene population growth, but this does not mean
that it happened in Africa. Humans are very mobile, and I don't
think we can assume that African ancestors were in Africa
50,000 years ago. Stan Ambrose and Richard Klein tell me that
there is so far little or nothing in the way of archaeological
remains from Africa between 75,000 and 25,000 years ago.
(This is apparently something that all archaeologists know but
none of the rest of us do.)

The available samples for our paper were far from
ideal. They were a total sample of whatever was available
scrounging in Mark Stoneking's lab at the time we wrote the
paper. But I think that our results show that it does not matter
very much for studying mtDNA. The reason is that most of the
differences between individuals in the world are about 50,000
years or so deep (i.e. the peak of our waves). Most ethnic
differences are much more shallow than this. In other words,
mtDNA is good for looking at the difference between
Europeans and Amerindians but almost worthless for looking
at the difference between, say, Italians and Swedes.

I agree that it is time to look for the staging areas of
the expansion of races. But remember that if these expansions
occurred from populations on the order of 10,000 total size
they could have been in pretty small restricted areas and hard to
find today. One candidate worth considering is the 82,000 year
old "Aurignacian" from the western rift in Zaire that Alison
Brooks and John Yellen are writing up.

John Relethford and I have been doing some work that
I want to mention to you because I want to throw the numbers
we are getting at you and see what you think the consequences
might have been for language evolution. We start with idea of
rather isolated precursor populations in their respective staging
areas and ask what the consequences would be for nuclear gene
frequency diversity and the accumulation of differences. We
then compare the theory with craniometric data collected by
Bill Howells in the late 1960s that he has graciously made
available to researchers. We chose three populations from each
of four major regions in the Old World — Europe, sub-Saharan
Africa, East Asia, and Australasia. The theory of all this is
impenetrable gobbledygook that makes the other paper looks
like child's play, and neither one of us really understands it all
in any deep sense, but it is amenable to the cookbook approach.
Our conclusions are (1) there was time for gene differences
among races to equilibrate in 20,000 or so years, after which
(2) population expansions would have "frozen" differences
among groups. This means that many of the race differences we
see today are quite easily accounted for by a model that says
they accumulated 100,000 years ago because of population
subdivision then.

The interesting conclusions are now (3) the ancestral
African population was three times as large as the ancestral
populations of Europeans, East Asians, or Australasians. That
is to say the ratios of long term effective sizes of these regions

were 3: 1: 1: 1, and (4) the rates of gene flow among regions
were as shown in the figure below. This is taken from our
submitted manuscript.

The numbers are the reconstructed number of
migrants per thousand years. These are "effective migrants" so
that old folks and kids would not count. These numbers are
invariant under changing assumptions about how large the
groups in the nesting areas were. In other words, craniometric
variation implies that there were 17 migrants per millennium
exchanged between ancestral Africans and East Asians no
matter what the total population of African and East Asian
ancestors was at the time. Note especially the implication that
there was almost no gene flow between northern and southern
Asians and that gene flow between African and European
ancestors was mostly via Asian ancestors rather than direct.
This pattern looks a lot like what Luca et al. come up with from
other marker frequencies.

I am pretty enthusiastic about this work, but we need
to repeat it on other traits like dermatoglyphics before we take
it too seriously. We have no idea in the world what the standard
errors of any of these numbers are. I expect that they are in the
right ballpark regardless of details. What would happen to
language here if one stray couple moved in every other
generation or so? My guess is that it would have no effect on
language differentiation in spite of its substantial effect on
genetic differentiation. If this intuition is right, then language
differences may be much older and deeper than genetic
differences.

Finally, let me recommend the best summary paper
about DNA and human evolution yet. Alan Rogers and Lynn
Jorde wrote it — they have submitted it somewhere, but I don't
know where. It is available through anonymous FTP from
anthro.utah.edu in /pub. I believe that the title is "origins" but
there are only a few papers there so it should be easy to spot.
There is a postscript version (.ps), a TeX version (.dvi), and an
HP version for direct printing to a laserjet (.hp).

Thanks again for the copy of Mother Tongue. Perhaps
I will see you at the Africanist Archaeology meetings in
Indiana?

Best regards,

Henry Harpending

Pennsylvania State University

Australasia

Europe

Africa

Far East

MOTHER TONGUE

Issue 22, May 1994

LETTER FROM YPSILANTI

Even being somewhat familiar with the harsh tone
often used in MT contributions, we were still surprised by Hal
Fleming's comments (MT 21, p. 68) on the Eastern Michigan
workshop on "Nostratic: Evidence & Status". We organized the
workshop intentionally as a small and informal working group,
one convened because some historical linguists sought a
scholarly dialogue about Nostratic between some practitioners
and some skeptics. While no one's major questions were
definitively answered, a number of us left with better
appreciation for some other positions, a sense of what kind of
evidence might confirm or disconfirm the hypothesis for us or
others and so forth.

Following the workshop, a volume of papers, co-
edited by Joseph & Salmons, was planned and is now in
preparation, to explore further questions raised in Ypsilanti.
We are aiming throughout to increase communication about
some intensely held views, thus serving the interests of
historical linguists of all ideological stripes, from the four
members of your organization listed on the inside cover of MT
whose work was presented (plus another who was invited but
unable to attend and yet another whose work will be in the
volume) to some very conservative Indo-Europeanists. This
kind of interaction among people of differing views seems to us
a normal part of scientific progress.

The ill-informed and insulting comments in MT only
give credence to views like those expressed by William Poser
(MT 21, p. 65) that a "cultish mentality... pervades... MT" and
that "Discourse that relies on ad hominem attack and team spirit
is unhealthy and unscientific." With him, we suggest that the
focus be kept on presenting solid data and substantive
argumentation, as we tried to do in Ypsilanti.

No doubt the writers in their irritation found it useful
to retaliate by repeating Poser's remarks. But that is a non
sequitur. Why do my inappropriate remarks show that Poser's
remarks were right?

Everyone ought to remember why Mother Tongue
exists in the first place and why its embattled efforts strike
some others as "cultish". When was the last time you saw a cult
that published all criticisms of itself? The real cult is the
establishment in historical linguistics. They treat their beliefs as
sacred dogma and do not allow free, open, and honest scientific
discussion. Otherwise we would have argued, as Poser suggests
we should, in journals like Language and UAL long ago. You
want us to stop speaking harshly? Fine, then get the editor of
Language to let us publish on a fair basis, tit for tat, in her
journal, instead of hiding behind "peer review" to keep us out,
while allowing harsh words about us to flow from the likes of
Rankin and Poser without any chance of our answering them.
Let us all wait and see if UAL has the scientific honesty to
publish Greenberg's long rebuttal to his critics. Then we can all
see if this dispute in the science of linguistics is a fair fight or
not. Surely, no one is surprised to find paradigm defenders
trying to control the flow of opinion or to squelch non-
conformists. It is apparently famous in the history of science,
although the innocents in the philosophy of science often seem
not to know this. So we must say to our critics: don't you think
there is just the faintest whiff of hypocrisy in the air?

ALL METHOD, NO CONTENT

HAROLD C. FLEMING
Gloucester, MA

Joseph C. Salmons

Alexis Manaster Ramer

Brian Joseph

Let this be the theme of Fleming's reply to the letter
from William Poser, which appeared in the previous issue of
Mother Tongue.

Organizing Committee,
Workshops on Comparative Linguistics

RESPONSE TO THE LETTER FROM
YPSILANTI

HAROLD C. FLEMING
Gloucester, MA

I ask the writers of the letter, complaining of my harsh
words about their Nostratic conference, to accept my apologies.
Truly! Sometimes I get carried away by the heat of combat with
the forces of ignorance. Then I say inappropriate things. My
remarks were inappropriate. Truly, I was responding to the
dismay of some colleagues who were not invited. But I was
wrong to criticize the conveners for having a small and select
group of participants.

Having successfully resisted the impulse to "come out
swinging with both fists", as they say in boxing, I decided that
this space was an excellent place to discuss things with our
opponents in a cooler way. Poser's letter (MT-21, p. 65) had a
hard polemic edge to it as has most of his writing attacking
Greenberg. It is awfully tempting to reply in kind. However,
his letter and his review of Ruhlen's hook truly showed the
themes of the opposition in a clear light. The full draft of the
letter which he and several other colleagues sent to the
Scientific American showed their position even more clearly. It
ought to be published in Mother Tongue with ample space
given for Greenberg to reply and others to comment as well.
More than half of the original signers of that letter are also
members of ASLIP. So here is a good opportunity to talk to
them too. At least one of them is showing signs of becoming
more reasonable.

Before going to the main theme, let me also respond
to the vital accusation of not having read Poser's article in UAL.
Actually, he can say that only by distorting what I said about
him in MT-19. It clearly did not depend on my having read his

MOTHER TONGUE

Issue 22, May 1994

article. Rather than making everyone go search for their copy
of MT-17 (p. 53), I will give the full quote here:

William Poser of Stanford mounts violent attacks on
Greenberg's Amerind from time to time and
apparently one of them got into UAL. Hearsay . (We
didn't get to UAL yet this year.) But it would be
splendid if one of our courageous Americanists
actually left the womb and counter attacked once in a
while. Naturally people would pass you in the
hallways without speaking and no one would invite
you to dinner anymore and life would be hell...

Indeed, I never read his article and probably never will
either. What I was reporting was the news given by a colleague
that Poser had attacked Greenberg in UAL. That news was
accurate and still is.

The main theme of Poser's MT letter and his review of
Ruhlen's book is that Greenberg and Ruhlen are
methodologically deficient and, therefore, their classifications,
their taxonomies, cannot be right. Poser and his fellows have
another theme, namely, that Greenberg's data are mistaken in
case after case and so, therefore, his taxonomy cannot be right.
His Iroquoian data are off, his Algonkian data are off, his
Slouan data are off, his Salinan data (presumably) are off, they
say. So how could he possibly be correct? However, we will
not deal with these Americanist particulars here, mostly
because Ruhlen or Greenberg himself in various places have
answered them already. (For a Ruhlen response to Rankin's
attack on Greenberg see Annual Review of Anthropology.
Greenberg has written a long rebuttal to his critics, which has
been submitted to UAL.)

What is most striking in Poser's letter and his review
of Ruhlen is the programmatic theme. Poser tells us how
classification is to be done. (Rankin in fact does much the
same.) Only after having avoided all the pitfalls of borrowing,
influence, etc., and having lined up sound correspondences, can
one begin to classify. If one cannot or will not do it that way,
then one can be attacked by a solid mainstream linguist as being
incompetent. (Poser calls Ruhlen "unreliable".) Besides this
being a classic Indo-European statement of how one begins
reconstruction, there is practically no successful classification
in the real world (including Indo-European) that has been done
this way. If our goal is not methodological conformity, it has to
be taxonomy, n'est-ce pas?

Where, one might ask, is the Amerind taxonomy
produced by Greenberg's opponents? They assure us that they
are the experts on native American languages, not Greenberg.
Well, then, where have they advanced the frontiers beyond
obvious (easy) taxa (classes)? It is not a rhetorical question: tell
me, where have they specifically done these things? Let the
baseline of comparison be the taxonomy of Lyle Campbell's
1979 book on North America and Loukotka's 1968
classification of South America. So what do we find? What
have they done? If we take Sapir's taxonomy as an earlier
baseline, have we not gone backwards since his time?

At least for now, I have to conclude that a bunch ot
idealists/programmers/perfectionists have told us that they
must be correct, even though they have not yet produced any

taxonomic advances, and that Greenberg must be wrong, even
though he has produced a major part of the taxonomy of the
entire world. Why would anyone want to abandon the schemes
of a proven taxonomist, indeed one of the most gifted
taxonomists to ever inhabit linguistics just to adopt the
proposals of unproven methodological theorizing?

Where do these proposals come from anyway?
Probably from refinements of Indo-European theory mixed
with the overwhelming emphasis on rigor and analysis of 20th
century linguistics. Why would anyone believe their proposals?
Probably because nearly all teachers of historical linguistics tell
them it is God's Own Truth and all righteous linguists should
follow the tenets given them by their teachers. Clearly,
Greenberg's most deplorable fault was not writing a popular
introductory text in Historical Linguistics.

Maybe this sounds too sociological? Maybe it really is
true that solid methodological theory is preferable to mere
historical hypotheses? Perhaps, but there would still be a
problem. Actually, two problems. First is that this standard or
textbook theory is not well supported empirically. People just
keep repeating it like the mantram it is and patting each other
on the back for understanding it. Falsification would be its true
destiny in a serious science, were it a true theory, but it is only
what I have called "guideline" theorizing (cf. the discussion of
dispersal theory in MT-21, p. 68). It adds up to a heuristic
device — and quite a good one — for reconstructing ancestral
languages. Read closely some of Henry Hoenigswald's
statements on "the comparative method"; note that he begins
with an existing taxonomy and proceeds to reconstruct the
ancestor of the languages in the taxon. Note that he does not set
up the etymologies either; nor does he judge their cogency.
That is natural because he thinks mostly of Indo-European, an
existing taxon where almost two centuries of discussion and
debate have established the good etymologies and weeded out
the bad.

Let one start out with Kikuyu, Masai, Luo, Boni, and
El Molo — all of Kenya — and see how far he gets with these
methods. It would be like using a microscope to find a boil on
a nose. (Three very distinct phyla are found in this Kenyan
group; Niger-Congo, Nilo-Saharan, Afroasiatic.)

The second problem in more general terms is the
imposition of some social science theory on history. Two
examples are given later. I submit that what Poser and his
colleagues are doing is imposing a scheme on Amerind data, a
scheme which is correct ideologically and professionally but a
scheme which basically prevents them from producing higher
level taxa. Should we all hang by our ears, waiting for the vast
Amazonian collection of languages to be reduced to even 10 or
20 genetic classes from its present 100+? Do not the opponents
of Greenberg feel any pressure to get a move on? Are they not
embarrassed that their New World lags far behind the rest of
the world — taxonomically?

Not everyone knows it, but some in Poser's camp
have been quietly searching for levers to pry loose the planks
of Greenberg's African classification. What better way to refute
Greenberg than to show that he was actually mistaken, and has
been mistaken for 30 years? But alas, they have had no success.
Perhaps it is helpful to repeat a few things that the critics cannot
seem to remember.

MOTHER TONGUE

Issue 22, May 1994

ONE is that Africa has far more languages than the
entire New World has, even including Eskimoan and Na-Dene
— about 1475+ in Africa to 625+ in the Americas.

TWO is that Africa has much more phonetic diversity
than the Americas or any other regions, even including the
Caucasus.

THREE, perhaps because of its morphological
conservatism, Africa has much typological diversity but
perhaps no more than, or even less than, the New World.

FOUR, in prominent or popular taxonomies of a few
generations ago Africa had only two or three overarching
"phyla"! These were the great typological schemes which gave
typology such a bad name in Africanistics. Greenberg actually
increased the African taxa to about 14 in 1953 before finally
reducing the whole lot to four phyla in 1963! His were genetic
classes, often with substantial typological dissimilarities within
them.

FIVE is that very little of the ordering of the vast
Africana was helped by reconstructions. A few existed in Bantu
and Semitic. To put it much more forcibly — reconstruction,
which is supposed to be necessary for classification beyond the
obvious, was irrelevant to the general outcome. And that went
far beyond the obvious. Indeed, some have argued that all
Greenberg did was to pull together the obvious taxa and put
them into four categories.

SIX is that the formal establishment of sound
correspondences was basically irrelevant too. I say "formal"
because the Greenberg brain was obviously seeing many
informal correspondences and using them. But, despite his
sophisticated brain, much of the classification did rest on plain
old similarity.

SEVEN is that Greenberg sometimes used poor data.
Sometimes the poor data, usually due to the fact that there was
only one source on the language, impeded things a little. Some
languages had to be put in their proper categories later on. But
mostly the data were not so bad in general and perhaps as a
whole not so much better nor worse than the Amerind data.

The EIGHTH and final thing is that, after thirty years
of the 4-phyla scheme and forty years of the 14-phyla scheme,
with lots of new languages being reported and eager young
scholars trimming and pruning the branches — the African
taxonomy is firm and strong. It isn't going to be moved
because it's the real thing!

It is natural for other anthropologists to see linguists
as myopic and small-minded — and indeed, many of them
really are. But we have the "my tribe", "my people" types in
social anthropology who sometimes know virtually nothing
except the incredibly interesting people they spent two years
with. Or the archeologist who keeps returning to the same site
year after year until he knows absolutely everything about very
little. These are sometimes the folks they have teaching the
introductory courses; luckily that work keeps them from
becoming completely myopic. Yet all of us yield to the
temptation of specializations, i.e., we think that our wee
endeavors are the model of what others ought to do.

Some things about Poser's constant attacks on
Greenberg and Ruhlen are deeply disturbing. His presumption

is enormous. His preconceived ideas govern everything he
writes about that I have seen. He dismisses Ruhlen's long hard
work on the global set of human languages, and he scorns
Greenberg's effort to bring a clear taxonomic picture to those
hundreds of native American languages, even when Greenberg
is still regarded by hundreds of social scientists as the unrivaled
master of linguistic classification in our whole history!

One would think that Poser and his gang would be
pleased at what Greenberg has done. One would think that they
would be out there testing all the connections between groups
and adding or subtracting etymologies — like scholars do in
fields where people are civil to each other. One wonders
whether Poser cares at all about taxonomy itself, the content of
classification, and all the history that implies. For sure, he never
seems to mention the specific content of Greenberg's Amerind
hypothesis. He reviewed Ruhlen's Guide without commenting
on the classification itself — what the book was about! Only
methodology, ever methodology, again methodology — yet his
has serious flaws in it!

We will let Cavalli-Sforza handle Poser's attack on
him, though he is more likely to ignore it. There is the matter of
the last pious statement. I will disregard the comment about
"cultish mentality" and that stuff. I find the last sentence rather
amusing — now quoting:

Only if they are able to present solid data and
substantive arguments will proponents of distant
genetic affiliation have an impact on the scientific
community.

Would Poser know what that stuff was if it came up and kissed
them? What do they think the Greenberg hypothesis is
anyway? There is no solid data? There are no substantive
arguments? Is it possible that Poser thinks that they alone
represent the entire scientific community? Rest assured, long
range hypotheses have already had an impact on the scientific
community and will have more. What Poser and his gang
should worry about is whether other scientists will begin to lose
respect for the dog-in-the-manger style of linguistics which he
represents so well.

AFTERTHOUGHT

There was a review of an interesting book on the
history and philosophy of science. Written by John H. Brooke
around 1987 and appearing in a journal whose name has been
lost, the review makes some sharp comments on the subject
dear to all of us. The editors are John A. Schuster and Richard
R. Yeo. The full title is given because it will be hard to find this
book: The Politics and Rhetoric of Scientific Method:
Historical Studies (Australasian Studies in History and
Philosphy of Science, 4), Dordrecht, Boston, and Lancaster: D.
Reidel, 1986. (Distributed in North American by Kluwer
Academic Publishers, Hingham, Mass.) Some quotes from the
review:

MOTHER TONGUE

Issue 22, May 1994

Do pronouncements on methodology, which
scientists habitually make, lose their interest once it is
appreciated that they rarely, if ever, determine
scientific practice? If, as with Newton's famous but
phoney "experimentum crucis", they have typically
fulfilled rhetorical functions, should they not be
disregarded? Emphatically not, is the message of this
book; for the rhetorical and political functions of
method discourse are precisely those that illuminate
the social dimensions of science. This is because
methodological prescriptions have assisted in the
presentation of controversial theories, in the
promotion of an individual's science as normative, in
accounting for the erroneous character of rival
science, in demarcation disputes between different
disciplines, and in cultivating images of scientific
unity likely to impress the public.

The eight case studies presented here
vindicate the editors' claim that a sociopolitical history
of methodology can prove extremely fertile...

However sterile formal methodologies may
be, they have undoubtedly been appropriated for
political ends ...

The assumption that there was a single,
accessible, and transferable scientific method was
commonly displayed to encourage recruitment to
scientific societies and to promote popular scientific
education...

Have methodological considerations never
impinged directly on the content of scientific theory?
Not all the contributors wish to go that far. Reporting
the controversial launch of "continental drift", Homer
Le Grand shows how Alfred Wegener adjusted the
presentation of his case in response to methodological
critiques ...

We will have more to say about Alfred Wegener in the future.

END NOTES

1) An example of this from a hostile British social
anthropologist: He told me that Greenberg had amazing insight
into the true forms of Hadza (Khoisan) words, not being fooled
by the inadequate description in the source used. "Somehow he
gets them right!", he said. Years later, at an international
conference, he denied ever having said that and denied that
Hadza was related to any other language in the world.
However, two of us had heard his earlier remark, and I had
recorded it in my notes.

1) One is from political sociology. A brilliant social
theorist from Harvard proposed an acute sociological analysis
of why Socialism never caught on in America. A historian
reviewed the analysis and said the whole idea was brilliant but,
unfortunately, false. Things had not happened that way in fact.

The second was a very neat sociological scheme
(theory) for analyzing the evolution of Ethiopia as a polity,
written by a friend of mine. As a reviewer, I had the
unfortunate duty of pointing out that the scheme was invalid;

things had not happened the way the scheme proposed.

These problems exist primarily because scholars
impose explanatory theory on the narrative of history but
without first establishing the narrative. This I derive from the
wisdom of a great philosopher of science, Karl Hempel, who
assumed that in the ordinary case the scientist would be dealing
with an established narrative, like European military history or
the dynasties of China. Then one cited the particular case
(narrative) as an example of the operation of a general law —
that explained it. While Hempel's theory is aging and needs
repair, the distinction is useful to make. For example, one may
want to explain the Russian, French, Ethiopian, and Mexican
revolutions. Then one must obtain the narratives of each
revolution but then one must have a general "Law of
Revolutions" to explain each or any one. (The Russian and
Ethiopian cases are fascinating in their similarity, both
eliminated old ruling classes and both acquired even worse
dictatorships.)

2) For non-linguists: We recommend that you treat the
term "typological" as usually equivalent to "grammatical",
meaning particularly morphology, syntax, and phonology. Two
languages may be highly similar typologically but not have the
same specific grammemes (bound forms usually) which can be
shown to be cognate. Thus the languages may not show any
evidence of genetic connection even when they are
typologically similar. And vice versa. German and English are
very close genetically but rather dissimilar typologically.

3) There is a fascinating parallel between the critics of
Greenberg's African hypotheses and the critics of his American
hypotheses. Once in Koln, I met a charming, bright, and well-
spoken Hungarian linguist, named Istvan Fodor. Istvan had
written a good-sized and rather impressive book attacking the
African classification — root and branch. Istvan showed how
Greenberg's methods could not possibly be correct, and he
attempted to show how wrong in linguistic details Greenberg
was, not overlooking the many data errors Greenberg had
made. Istvan's book was designed simply to blow Greenberg
and his whole approach right out of the water. Critics of the
Amerind hypothesis should go look up Fodor's guide to
African misclassification. But they should also note that Istvan
Fodor had almost no impact at all on the final outcome. Perhaps
a little in Hungary or maybe in Germany, because he was in
Koln. A good classification can withstand any number of
methodological criticisms. Why? This may sound corny, but
probably a good classification is strong because it is true.

4) For example, some of our long rangers belong to
the Cherokee nation. They can wonder about their roots just as
much as an African American or a member of the "first families
of Virginia" (FFV) can. They can get very little satisfaction
from finding out that their roots go to Iroquoian and then
maybe a little deeper to some connection with Siouan. But the
traditional Americanist can only tell them that their roots are
unknown, since properly prestigious scientific linguistics can
only go back a few thousand years. Good thing our Cherokees
have archeologists around to tell them they go back at least to
the Clovis horizon or biogeneticists to tell them that their
ancestors ultimately came from eastern Asia as long as 20 to 30
millennia ago.

MOTHER TONGUE

Issue 22, May 1994

ASLIP BUSINESS

Our annual meeting was held in Boston at the African
Studies Center of Boston University on 16 April 1994. It was a
great success in many respects, perhaps most of all: good
fellowship.

The formal elections produced the following results
(addresses are shown on the inside of the front cover of this
newsletter):

President:
Vice President:
Secretary:

Harold C. Fleming
Allan R. Bomhard
Anne W. Beaman

The President also serves as Treasurer. All
correspondence about membership dues, orders for back issues,
and the like should be addressed to him. (Some people have
been uncertain about this.)

The following colleagues were formally elected to the
Board of Directors (addresses are not given but can be obtained
by inquiring to the President):

Ofer Bar-Yosef
Ronald Christensen
Frederick Gamst
John Hutchison
Mark Kaiser
Mary Ellen Lepionka
Philip Lieberman
Daniel McCall
Roger Wescott

Other Business of the Annual Meeting:

1) On the question of what to do concerning a member
of the Council of Fellows because of his/her non-
communication for very long periods of time and non-renewal
of membership. The Board strongly disagreed with the
President, believing that such a removal would serve no
purpose. The Board rules ASLIP; hence an elected Fellow is
elected for life, no matter what s/he does to or for us. The value
of honoring Fellows is higher than any functional value a
Fellow may have to ASLIP. We come to praise them, not use
them.

2) On the date of the Annual Meeting: A member of
ASLIP wished to have the By Laws changed so that the Annual
Meeting could be held in March. That motion was voted down.
In place of the existing wording, the following was substituted:
"The Annual Meeting will be held on or about April 15th,
preferably on the first Saturday after the 15th."

3) On the question of the nature of the Board and the
Annual Meeting, the President proposed the following, which
the Board and other members present did not challenge or
disagree with:

a) The Board should be likened to a Board of Trustees,
rather than to an executive group. Its primary duty is
to preserve ASLIP and its values and purposes and to
prevent either the officers or the editor of Mother
Tongue from subverting those values and purposes.
The Board can be likened to a gyroscope.

b) It is not necessary for the Directors to be linguists or
physical anthropologists or archeologists or even to
be intensely interested in the newsletter. Like the
trustees of a university, they can be drawn from many
careers. But their basic moral commitment to the
good of the organization is paramount.

c) The Annual Meeting is primarily a meeting of the
Board of Directors and the Officers, who perform
their duties annually in conformity to the By Laws.
Hence the Annual Meeting is largely a ceremonial or
ritual occasion, although it can discuss anything new
that it wants to and may dispute old decisions.

d) All power or final power in ASLIP resides in the
membership. Were a large enough contingent of
members to attend the Annual Meeting they could
(theoretically) remove the current members of the
Board and the current officers and choose ones more
to their liking. That they have not done so is
instructive.

e) Even though the time and place of the Annual
Meeting have been clearly advertised in Mother
Tongue for at least 5 years, no members from outside
of New England have come to the meetings, until this
year. Suggestions for annual meetings or occasional
meetings at times and places other than the traditional
ones have been met with indifference. The reasons
for that are probably obvious to most of us.

f) Therefore, the President proposed that we draw the
conclusion that we are primarily a group of readers,
not conveners. Our older larger counterpart is
Current Anthropology, rather than the AAA or LSA
or S AA or ASPA. Whatever desires we may have had
to be socially interactive are gone. Power lies in the
membership, and the membership has clearly defined
ASLIP as a publication.

g) So, though we may call some conferences at some
points on some topics in trying to act out our basic
charter, it will not be a regular or annual feature of
ASLIP. We are essentially out of the convention
business.

MOTHER TONGUE

Issue 22, May 1994

EDITORIAL

MISCELLANEA

In this editorial, I would like to lay out some
guidelines for the preparation of papers to be published in
Mother Tongue.

General: Mother Tongue aims to appeal to a wide
audience: linguists, anthropologists, archeologists, and
geneticists, among others. Keeping this in mind, authors
should avoid the use of technical jargon as much as possible. If
it is necessary to use technical jargon, it should be carefully
defined. Moreover, many of our readers are not native
speakers of English. Therefore, authors should avoid using
colloquial expressions. As a general rule, keep it simple, keep
it clear, and keep it concise. All papers must be in English.
Finally, try to keep the rhetoric civil. I can attest from personal
experience, having myself been the target of ad hominem
attacks in earlier issues of this newsletter, that it is not pleasant.

Cited Forms: Please try to avoid idiosyncratic
symbols. If it is necessary to use special symbols, define them.
For example, who, beyond a narrow group of specialists, would
know that the symbol /y/ represents a voiced alveolar lateral

fricative, which is usually represented by the symbol /fe/? If in

doubt about acceptable symbols, you may want to consult the
Phonetic Symbol Guide by Geoffrey K. Pullum and William A.
Ladusaw (Chicago, IL: University of Chicago Press, 1986) or
Patterns of Sounds by Ian Maddieson (Cambridge: Cambridge
University Press, 1984), both of which should be relatively
easy to obtain. If you are trying to make a point and are using
cited forms to illustrate your arguments, you will fail to
convince if nobody can understand your examples because you
have used idiosyncratic symbols. Or, refer to a standard (for
example, Africanist, Indologist, etc.) so that others can translate
your symbols into theirs. Best of all, create a table of
equivalents to show just how your symbols match some known
standard. Finally, all cited forms should be followed by an
English gloss, thus: Egyptian dd n-fnswt 'the king speaks to
him', Proto-Indo-European *medhio- 'middle', Monguor
fsiargi- 'to swallow', Svan mi T. Do not use 'id.'

Abbreviations: Beyond a few common
abbreviations such as "mtDNA", "etc.", "et al.", for example,
abbreviations should be avoided. This also applies to the
names of languages and dialects, which should be spelled out
in full. If you must use abbreviations, define them.

References: Bibliographical references should be
as complete as possible. For format, we suggest that you
follow the Chicago Manual of Style (14th edition, 1993),
chapter 16 — note especially figures 16.1 and 16.2 on pages
647 and 648. Again, avoid abbreviations — spell out the titles
of articles, books, and journals fully. Give page numbers. This
does not apply to letters or Quick Notes, of course.

Allan R. Bomhard

NEW BOOKS

The Origin of Language: Tracing the Evolution of the
Mother Tongue. By Merritt Ruhlen. New York, NY: John
Wiley & Sons, 1994, pp. xi, 239. $27.95.

"Just as archeologists study fossils and ancient artifacts for
clues about mankind's origins, linguistic researchers today are
sifting through word roots and grammatical conventions and
coming up with startling revelations about our beginnings. In
The Origin of Language noted linguist Merritt Ruhlen takes
you on a fascinating journey of discovery back through nearly
100,000 years of human history and prehistory in pursuit of the
language from which all modern tongues derive.

"Requiring no prior familiarity with linguistics, The Origin of
Language is the first book to explain, in laymen's terms, the
controversial process by which linguists are tracing the
development of the vast range of human speech, sweeping
aside many traditional assumptions about the spread of
language and the roots of the human family. In addition to
acquainting you with the manner in which such diverse
languages as English and Chinese can be compared, Dr. Ruhlen
introduces you to the brilliant mavericks whose linguistic
theories are at last gaining worldwide acceptance. He also
discusses the exciting new work being done in genetics and
archeology that corroborates much of the controversial
evidence.

"But more than simply describing his and his colleague's
theories, Dr. Ruhlen invites you to share in the joys of
discovery. He arms you with the linguist's basic toolkit and
lets you work through the evidence for yourself and draw your
own conclusions. You'll classify languages and language
families, trace language family trees, and even reconstruct
some of the basic vocabulary used by our most distant
ancestors. Also, based on clues provided by your research,
you'll plot the land and sea routes most likely taken by early
humans in their diaspora out of Africa and to the four corners
of the world.

"While the Origin of Language is an incomparable introduction
to some of the most exciting linguistic research now being
conducted by researchers around the globe, it is also much
more. It is an inspiring invitation to join the quest for our
human roots and to hear echoes of the Mother Tongue."

The Nostratic Macrofamily: A Study in Distant
Linguistic Relationship. By Allan R. Bomhard and John
C. Kerns. (Trends in Linguistics, Studies and Monographs 74.)
Berlin and New York, NY: Mouton de Gruyter, 1994, pp. xi,
932. $199.00 (DM 298.00).

MOTHER TONGUE Issue 22, May 1994

In this book, the authors present extensive evidence that certain
languages/language families of Europe, Asia, and northern
Africa are remotely genetically related. First, they agree with
Greenberg that Indo-European belongs, along with Uralic-
Yukaghir, Altaic (Mongolian, Manchu-Tungus, Chuvash-
Turkic, Korean, and probably Japanese-Ryukyuan), Chukchi-
Kamchatkan, Gilyak, and Eskimo-Aleut, to the Eurasiatic
language family. Then, they maintain that Eurasiatic, as a
group, is related to Elamo-Dravidian, Kartvelian, and, more
distantly, Afroasiatic, all of which together constitute the
Nostratic macrofamily.

Bomhard and Kerns begin by surveying the Nostratic
languages. Next, they discuss the phonological systems of the
Nostratic daughter languages, ending with a table of
correspondences. Then, they trace the Indo-European
phonological system through various stages of development.
The following chapter is devoted to an overview of Nostratic
morphology. Finally, the core of the book, running well over
500 pages, is the presentation of 601 Nostratic etymologies,
supported by voluminous data from the various Nostratic
daughter languages.

FONTS

Allan Bomhard has developed the Phonetic Symbols Character
Set for use with MicroSoft Windows 3.1. (these are the phonetic
symbols used in preparing Mother Tongue). The Phonetic
Symbols Character Set contains over 800 fully scaleable True
Type special phonetic symbols in the popular Times Roman
typeface. Regular, italic, bold, and bold italic members are
included. The phonetic symbols are spread across five font
files, the first two of which contain vowel symbols, and the
remaining three of which contain stop, fricative, and affricate
symbols. Order follows that of the Latin alphabet. An IBM or
compatible PC capable of running MicroSoft Windows 3.1 with
at least 1.5 Mb of free hard disk space are required. These
phonetic symbols are available for $20.00 (U.S. and Canada —
$25.00 elsewhere) (the fonts are essentially being made
available at no charge — the $20.00 is to cover the cost of
postage, media, installation instructions, mailing envelope,
etc.). Checks should be made payable to "Allan R. Bomhard".
Orders may be placed with (please indicate preference: 3.5
inch or 5.25 inch diskette):

Allan R. Bomhard

73 Phillips Street

Boston, MA 021 14 U.S.A.

OTHER TONGUE

NEWSLETTER OF THE ASSOCIATION FOR
THE STUDY OF LANGUAGE IN PREHISTORY

AIM AND SCOPE

The Association for the Study of Language in Prehistory (ASLIP) is a nonprofit organization,
incorporated under the laws of the Commonwealth of Massachusetts. Its purpose is to encourage
and support the study of language in prehistory in all fields and by all means, including research on
the early evolution of human language, supporting conferences, setting up a databank, and
publishing a newsletter and/or journal to report these activities.

MEMBERSHIP AND SUBSCRIPTION INFORMATION

Annual dues for ASLIP membership and subscription to Mother Tongue are US $15.00 in all countries except those with
currency problems (in those countries, annual dues are zero). Checks should be made payable to "ASLIP" and sent to:

Harold C. Fleming, President

Association for the Study of Language in Prehistory

16 Butman Avenue

Gloucester, MA 01930 U.S.A.

European distribution: All members living in Europe (up to the borders of Asia), and not having currency problems, should
pay their annual dues to and will receive Mother Tongue from:

Prof. Dr. Ekkehard Wolff

Universitat Hamburg

Seminar fur Afrikanische Sprachen und Kulturen

Rothenbaumchaussee 67/69

20148 Hamburg

Federal Republic of GERMANY

Payment must accompany all orders (except as noted above). Selected back issues of Mother Tongue are available for US
$4.00 each — for more information or to place an order, write Harold Fleming at the above address.

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Texts

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Mother Tongue Newsletters"

See other formats